Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author.
Ecology, epidemiology and evolution of enteric
microbes in fragmented populations of the endangered
takahe (Porphyrio hochstetteri)
A thesis presented in partial fulfilment of the
requirements for the degree of
Doctor of Philosophy
in
Veterinary science
at
Massey University, Manawatū,
New Zealand
Zoë Lorraine Grange
2015
I
Abstract
Pathogenic diseases are increasingly recognised as a challenge to the conservation of wildlife.
Complex host-pathogen relationships and transmission dynamics in wild populations can limit our
understanding of how pathogens contribute to the decline and endangerment of wildlife. Endangered
wildlife populations maintained in reserves present a unique opportunity to investigate wildlife host-
microbe relationships in a controlled semi-natural environment where diversity, abundance and the
movement of species are restricted. The aim of this study was to investigate the prevalence and
molecular differentiation of enteric bacteria carried by endangered takahe (Porphyrio hochstetteri).
Through the use of network analysis and molecular epidemiology, the study explored the effects of
geographic isolation and translocation on the prevalence, transmission and evolution of
Campylobacter and Salmonella spp. within fragmented populations of takahe.
Translocation and conservation management has created a dynamic network of takahe populations
which vary in their likelihood to maintain and transmit pathogens. My study suggests that range
expansion following a significant bottleneck and intensive conservation management of takahe has
had unforeseen consequences on microbial diversity. The management of takahe in different
environmental settings has influenced the carriage of Campylobacter jejuni and Campylobacter coli.
A newly discovered rail-associated Campylobacter sp. nova 1 was prevalent in all populations.
However, more discriminatory whole genome analysis of isolates detected a significant biogeographic
variation in C. sp. nova 1 genotypes. Possible explanations for the observed pattern include the spatial
expansion and isolation of hosts resulting in reduced gene flow of Campylobacter spp. and allopatric
speciation, and the presence of heterogeneous environmental attributes or cross-species transmission
of Campylobacter spp. from sympatric reservoir hosts. An assessment of vertebrate reservoirs in an
island ecosystem indicated cross-species transmission of Campylobacter spp. was not likely to be a
factor contributing to the maintenance and phylogeographical distribution of Campylobacter spp. in
takahe.
This study was the first of its kind to explore microbial dynamics in a large proportion of a well-
described but fragmented population of a wild bird. Results suggest historic and current management
practices may be having unforeseen influences on enteric microbes, the consequences of which are
unknown but could be detrimental to the health of translocated populations of takahe.
II
Acknowledgements
This project has been one of the most memorable experiences of my life. I have had the privilege of
working with the unique fauna of New Zealand within locations I could have only dreamt of. I would
like to thank the Department of Conservation and the Maori community for supporting my study and
giving me the privilege to work with, in my opinion, with the most beautiful bird in New Zealand, the
takahe. I would like to thank a few people within the Department of Conservation personally; Linda
and Chris Birmingham, Kate McInnes, Phil Marsh and Glen Greeves have all had important roles in
the organisation and practical aspects of this thesis.
To my primary supervisor Brett Gartrell, I hope the journey hasn’t been too tough. Thank you for
propping me up when I needed it and being a sounding board when my mind was in a muddle. I
couldn’t have asked for a better mentor and just because this thesis is complete, it doesn’t mean you
have seen the end of me.
I certainly did not complete this thesis alone; I was fortunate enough to have a prime selection of co-
supervisors to provide mentorship and who were kind enough to share their wisdom along the way.
To Nigel French, even though I did go on a helicopter a few times during this project, I promise you
this is not my normal mode of transport. I am forever grateful to you for providing a supportive and
caring environment to work in. I thank you for giving me opportunities to shine and work at the
cutting edge of science. To Patrick Biggs, you must have the patience of a saint to teach an ecologist
databases and genomics. Thank you for your guidance, support and colour expertise. I dedicate the
beauty of the figures contained in this thesis to you. To Nicky Nelson and Laryssa Howe, your
insights have helped make this thesis what it is and I am proud to have had you on my supervisory
panel.
I extend my gratitude to the collaborators I have worked with in the formation of this thesis. I was
lucky enough to gain a treasured friend in Mary van Andel, and I thank her for her endless patience.
Additionally, much appreciation goes to Jonathan Marshall and Marti Anderson for statistic
knowledge and advice.
Moving half way across the world has been made easy due to the wonderful people I have met during
this chapter of my life. There are many who have welcomed me into their lives along the way and to
whom I am thankful for their support, friendship and laughter. A special mention must be for Micah
Jensen and Aditi Sriram. You have seen me at my best and worst, you provided love, support and
III
laughter when I needed it most, and I could never have wished to meet such kind and thoughtful
friends.
It is no secret that I have moved a few times during this PhD. Each and every house has been a home
away from home and for this reason I thank Nicky Denning, Aditi Sriram, Dirk Steenbergen, and
Graeme and Serena Finlayson for your warmth and kindness.
Completing a PhD would never be possible without fellow students to share the highs, the lows and
somewhere in between. To my proof reader and kind friend, Springer Browne, you are one of the few
to have or will ever read this entire thesis word for word. I am forever indebted to you for your
grammatical abilities, although I feel payment in wine may be an option. Kruno Bojanic, you made
me smile every day and my life is brighter with you in it. Anja Freidrich, your company has kept me
entertained and motivated to continue throughout the years, especially during the final stages. To Kyle
Richardson, I hereby disclose this thesis as proof that I am an ecologist, no more evidence required.
The beauty of my fieldwork was that I was able to share it with other people whom were just as
passionate about wildlife and crazy enough to work from dusk to dawn. I was fortunate enough to
take along friends and colleagues including Thomas Burns, Micah Jensen, Sarah Michael, Danielle
Sijbranda, Pauline Nijman and Brett Gartrell. The unique memories we share of our times on Maud
Island are one of the main reasons I love what I do.
Angie Reynolds, without you I am sure I would still be in the lab to this day. Your kindness and
dedication to help is unmatchable. To Anne Midwinter, I am grateful for your emotional support and
enduring my ‘Englishness’ for the last three years. You certainly won the bet but I was close, surely I
deserve one bottle of champagne.
To the team at Wildbase, you have been my ongoing source of entertainment and support. Thank you
for taking me under your wing, teaching me the veterinary aspects of wildlife health and being the
source of my likeminded friends.
This thesis would never have been what it is without the financial backing of the Allan Wilson Centre.
I am proud to be associated with a collaborative institute which encourages the development of
scientists. Additionally, I would like to thank Massey University and the Institute of Veterinary and
Biomedical sciences for financial assistance to attend conferences far and wide in order to
communicate my work to a large audience and gain insight from my peers. Additionally, I am grateful
to the university avian health grant for providing funds towards my research.
IV
I would like to acknowledge the three examiners of this thesis, Daniel Thompkins, Wendi Roe and
Jonas Waldenstrom for their time and patience in reading this thesis word for word and providing
useful insights into my research.
Finally, I dedicate this thesis to my parents, Robin and Janey, my sisters, Laura and Kelly, and my
wonderful nephews, Max and Nicholas, to whom I have been a virtual daughter, sister and auntie for
the last 3 years. Without your love and support, this story would never been told, or not by me
anyway.
V
Thesis structure and format
This thesis is presented as a series of seven chapters. Encompassed by a general introduction and
discussion, five research chapters have been prepared and are presented as discrete papers for
publication in peer reviewed journals.
Chapter one
General introduction introduces the concepts behind the research contained in this thesis by
discussing and reviewing current literature on the principles of disease ecology and aspects of the
epidemiology of infectious organisms in wildlife. The objectives of the study are summarised at the
end of this chapter.
Chapter two
Network analysis of translocated takahe populations to identify disease surveillance targets has
been published in the journal Conservation Biology (Grange et al. 2014).
Chapter three
Using a common commensal bacterium in endangered takahe (Porphyrio hochstetteri), as a
model agent to explore pathogen dynamics in isolated wildlife populations is in press in the
journal Conservation Biology.
Chapter four
Wildlife translocation and the evolution and population structure of a host associated
commensal Campylobacter spp. is under review in the journal Proceedings of the National Academy
of Sciences (PNAS) following publication of chapter 3.
Chapter five
Investigation of vertebrate reservoirs of Campylobacter spp. in an island ecosystem will be
submitted to the Journal of Animal Ecology pending publication of chapters 3 and 4.
VI
Chapter six
Location specific prevalence of Salmonella spp. in endangered takahe (Porphyrio hochstetteri)
will be submitted to the Journal of Wildlife Disease.
Chapter seven
General discussion summarises the significant findings of this study. The relevance and implications
of results are discussed and future research directions are suggested.
Chapter eight
Literature cited has been collated at the end of the thesis to reduce repetition. Literature is referred to
in the format consistent with the format used for the journal Conservation Biology.
Chapter nine
Appendix contains all supplementary information organised by chapter
VII
Table of Contents
1. General introduction ........................................................................................................... 3
1.1. Wildlife disease ecology ............................................................................................. 3
1.1.1. Disease ecology concepts ................................................................................................ 3
1.1.2. Host pathogen relationships: from individuals to ecosystems ........................................ 4
1.1.3. The host population ......................................................................................................... 6
1.1.4. Multi-host pathogens and reservoir dynamics ................................................................ 7
1.2. Conservation of wildlife in the face of disease threats................................................ 9
1.2.1. Disease threats posed to and from wildlife ..................................................................... 9
1.2.2. Anthropogenic management of threatened wildlife populations: translocations and
sanctuaries as conservation tools .................................................................................................. 12
1.3. Epidemiological tools for species conservation ........................................................ 15
1.3.1. Risk assessments and disease surveillance ................................................................... 15
1.3.2. Pathogen ecology and epidemiology ............................................................................ 18
1.4. Research focus........................................................................................................... 19
1.4.1. New Zealand conservation management ...................................................................... 19
1.4.2. Takahe (Porphyrio hochstetteri) ................................................................................... 21
1.5. Microbes of interest to this study .............................................................................. 24
1.5.1. Campylobacter species ................................................................................................. 24
1.5.2. Salmonella species ........................................................................................................ 25
1.6. Objectives of the study .............................................................................................. 26
2. Network analysis of translocated takahe populations to identify disease surveillance
targets ....................................................................................................................................... 29
2.1. Abstract ..................................................................................................................... 29
2.2. Introduction ............................................................................................................... 30
2.3. Methods ..................................................................................................................... 32
2.3.1. Data set .......................................................................................................................... 32
2.3.2. Network description and topology ................................................................................ 33
VIII
2.3.3. Network dynamics ........................................................................................................ 34
2.4. Results ....................................................................................................................... 35
2.4.1. Network description and topology ................................................................................ 35
2.4.2. Network dynamics and node-level analysis .................................................................. 35
2.5. Discussion ................................................................................................................. 40
2.5.1. Application of network analysis to takahe movements ................................................. 40
2.5.2. Identification of hubs, sinks, and sources ..................................................................... 41
2.5.3. Limitations .................................................................................................................... 43
2.5.4. Conservation implications and future directions ........................................................... 44
2.6. Acknowledgments ..................................................................................................... 46
2.7. Supporting Information ............................................................................................. 46
3. Using a common commensal bacterium in endangered takahe (Porphyrio hochstetteri),
as a model to explore pathogen dynamics in isolated wildlife populations ............................. 51
3.1. Abstract ..................................................................................................................... 51
3.2. Introduction ............................................................................................................... 52
3.3. Materials and methods .............................................................................................. 53
3.3.1. Study population ........................................................................................................... 53
3.3.2. Sample collection .......................................................................................................... 53
3.3.3. Microbiological culture and DNA extraction ............................................................... 54
3.3.4. Molecular confirmation and speciation ......................................................................... 54
3.3.5. Prevalence of Campylobacter spp. in takahe ................................................................ 55
3.3.6. Exploratory analysis of explanatory covariates ............................................................ 56
3.3.7. Multiple correspondence analysis ................................................................................. 58
3.3.8. Multivariate logistic regression modelling .................................................................... 58
3.4. Results ....................................................................................................................... 59
3.4.1. Apparent prevalence of Campylobacter spp. in takahe................................................. 59
3.4.2. Estimates of true prevalence using imperfect tests ....................................................... 60
3.4.3. Analysis of explanatory covariates associated with the carriage of Campylobacter spp.
61
IX
3.5. Discussion ................................................................................................................. 64
3.6. Acknowledgements ................................................................................................... 67
3.7. Supporting information ............................................................................................. 68
4. Wildlife translocation and the evolution and population structure of a host-associated
commensal Campylobacter spp. .............................................................................................. 71
4.1. Abstract ..................................................................................................................... 71
4.2. Introduction ............................................................................................................... 72
4.3. Methods ..................................................................................................................... 74
4.3.1. Sample collection and culture ....................................................................................... 74
4.3.2. Selection of C. sp. nova 1 for genomic sequencing ...................................................... 74
4.3.3. Genomic DNA preparation and processing .................................................................. 74
4.3.4. Genome assembly, curation and annotation .................................................................. 75
4.3.5. Ribosomal multi locus sequence typing (rMLST) of C. sp. nova 1 .............................. 75
4.3.6. Core genome and rMLST tree construction .................................................................. 76
4.3.7. Multivariate analysis of the relationship between location factors and genetic distance
77
4.4. Results ....................................................................................................................... 77
4.4.1. C. sp. nova 1 comparison to published Campylobacter spp. ........................................ 77
4.4.2. Genomic differentiation of C. sp. nova 1 isolates ......................................................... 81
4.4.3. Multivariate analysis of C. sp. nova 1 rMLST allelic profiles ...................................... 81
4.5. Discussion ................................................................................................................. 84
4.6. Acknowledgements ................................................................................................... 88
4.7. Supporting information ............................................................................................. 89
5. Investigation of vertebrate reservoirs of Campylobacter spp. in an island ecosystem ..... 93
5.1. Abstract ..................................................................................................................... 93
5.2. Introduction ............................................................................................................... 94
5.3. Methods ..................................................................................................................... 95
5.3.1. Study site ....................................................................................................................... 95
X
5.3.2. Study populations .......................................................................................................... 96
5.3.3. Sample collection .......................................................................................................... 97
5.3.4. Microbiological culture, molecular confirmation and speciation ................................. 97
5.3.5. rMLST analysis ............................................................................................................. 98
5.3.6. In silico PCR of the C4-dicarboxylate trans-membrane transport gene ........................ 98
5.4. Results ....................................................................................................................... 98
5.4.1. Prevalence of Campylobacter spp. in vertebrate communities ..................................... 98
5.4.2. Comparative genomics of Campylobacter spp. ............................................................ 99
5.1. Discussion ............................................................................................................... 103
5.2. Acknowledgements ................................................................................................. 107
5.3. Animal ethics and permits ....................................................................................... 107
5.4. Supplementary information ..................................................................................... 107
6. Location specific prevalence of Salmonella spp. in endangered takahe (Porphyrio
hochstetteri) ........................................................................................................................... 111
6.1. Abstract ................................................................................................................... 111
6.2. Introduction ............................................................................................................. 112
6.3. Methods ................................................................................................................... 113
6.4. Results ..................................................................................................................... 114
6.5. Discussion ............................................................................................................... 115
6.6. Acknowledgements ................................................................................................. 117
7. General Discussion ......................................................................................................... 121
7.1. Microbial dynamics in translocated takahe (Porphyrio hochstetteri) ..................... 121
7.2. Disease risks associated with translocations ........................................................... 123
7.3. Advancing tools for epidemiological investigations of wildlife ............................. 126
7.4. Implications for conservation management ............................................................ 127
7.5. Future research directions ....................................................................................... 129
7.6. Concluding remarks ................................................................................................ 131
8. Literature Cited ............................................................................................................... 135
XI
9. Appendix ........................................................................................................................ 161
9.2. Chapter 2 supplementary information ..................................................................... 161
9.3. Chapter 3 supplementary information ..................................................................... 167
9.4. Chapter 4 supplementary information ..................................................................... 179
9.5. Chapter 5 supplementary information ..................................................................... 184
XII
List of tables and figures
Tables
Table 2-1 Takahe network measures ...…………………………………………………………36
Table 2-2 Network key locations....……..…...…………………………………………………38
Table 3-1 Takahe sampling effort and variables…..……………………………………………57
Table 3-2 Multivariate models for Campylobacter spp. carriage in takahe ……………………63
Table 4-1 Campylobacter sp. nova 1 PERMANOVA models.....………………………………82
Table 5-1 List of hosts and Campylobacter spp. prevalence on Maud island….…………..…100
Table 5-2 rMLST allelic profiles of Campylobacter spp. on Maud Island …..…………….…102
Table 6-1 Apparent prevalence of Salmonella spp. in takahe populations……………………114
Table 6-2 True prevalence of Salmonella spp. in takahe by sample type…..…………………115
Figures
Figure 1-1 Theoretical hypotheses for the modes of speciation………………..…………………5
Figure 1-2 Pathogen transmission dynamics between native and introduced populations…...…15
Figure 1-3 Takahe..………………………………………………………………………………21
Figure 1-4 Map of takahe distribution in New Zealand ...………………………………………23
Figure 2-1 Takahe translocation networks ...……………………………………………………37
Figure 2-2 Relationship between in degree and out degree network measures …………………39
Figure 3-1 Apparent and true prevalence of Campylobacter spp. in takahe.……………………60
Figure 3-2 Venn diagram of Campylobacter spp. carriage in takahe……………………………61
Figure 4-1 Map of sampling locations...…………………………………………………………78
Figure 4-2 Takahe Campylobacter spp. rMLST tree….…………………………...……………79
Figure 4-3 Takahe Campylobacter sp. nova 1 core and rMLST trees.…..…………………...…80
Figure 4-4 Takahe Campylobacter sp. nova 1 FST tree……………….………….……...………83
Figure 4-5 Schematic of hypotheses for Campylobacter sp. nova 1 genotypes in takahe………86
Figure 5-1 Maud island Campylobacter spp. rMLST tree and distance matrix……………..…101
XIII
CHAPTER 1
GENERAL INTRODUCTION
1
2
1. General introduction
1.1. Wildlife disease ecology
1.1.1. Disease ecology concepts
The impact of infectious disease on the global biodiversity of wildlife and the emergence of human
pathogens of zoonotic origins are becoming increasingly apparent as areas of concern. Although not a
common driver of extinction (Heard et al. 2013; Smith et al. 2006), infectious disease is a threatening
process which may impact species conservation (Heard et al. 2013; Smith et al. 2009). Catastrophic or
chronic host population depression may occur as a result of pathogenic outbreaks due to pathogen
induced death, increased susceptibility to predation and reduced reproductive success (Boadella et al.
2011; Cunningham 1996). Disease is defined as an abnormal condition of an organism, which can be
caused by infection with a pathogen. The terms “parasite”, “pathogen” and “infectious agent” are
often used interchangeably. In terms of transmission, pathogens transmit to new hosts not diseases. A
pathogen is any infectious organism that causes disease symptoms in its host. Disease ecology is the
study of interactions between hosts and pathogens, including variation in infection, transmission and
impacts of pathogens on host populations (Archie et al. 2009; Kilpatrick & Altizer 2010).
Exposure of a host to a pathogen and establishment of infection is the result of complex ecological
and evolutionary interactions between the host, the pathogen and the environment or habitat of the
host and pathogen, sometimes termed the epidemiological triangle (Silvy 2012; Wobeser 2006).
However, investigations of wildlife disease are often more complex than the epidemiological triangle
suggests, with underlying hierarchical structure and heterogeneous interactions in natural ecosystems
adding an additional level of complexity to the transmission and maintenance of infectious organisms
(Caillaud et al. 2013). Management and mitigation of wildlife disease is enhanced by an
understanding of the ecological factors influencing the transmission and interspecific variation of
microorganisms at the level of the host through to the population, community and ecosystem
(Tompkins et al. 2011).
3
1.1.2. Host pathogen relationships: from individuals to ecosystems
1.1.2.1. The host and the pathogen
The functional dependency of a pathogen on a host leads to an inevitable close association between
host and pathogen dynamics. Transmission to and colonisation of new hosts is required for the
persistence of infectious organisms in an ecosystem. Microorganisms are an important component of
biodiversity, contributing to ecosystem function and dynamics. A microorganism which is dependent
on a host for its survival may enter into one of four types of relationships: (i) commensalism, the
microorganism benefits without detrimental effects on the host, (ii) mutualism, both the host and
microorganism experience increased fitness, (iii) parasitism, the microorganism negatively impacts its
host’s fitness and (iv) opportunism which is a subset of parasitism where microbes whom are
commensal in healthy hosts may cause disease in compromised hosts (Goering & Mims 2013).
Therefore, not all infections result in the expression of disease and the nature of the association
between a host and an infectious organism may be dependent on a range of factors relating to the host
and / or pathogen. Colonisation of a host is dependent on the likelihood of transmission and host
defences such as behaviour and immunity. Multiple forms of pathogen transmission have been
described including: direct contact, indirect contact, or by vectors. Many biotic and abiotic factors
may contribute to differences in pathogen exposure during a host’s lifetime. Such factors include: age,
sex, habitat selection, population density, diet, social structure and the behaviour of the host (Johnson
et al. 2012).
Different lifestyles of pathogens are expected to influence their genomic composition. For example,
local environmental and host associated factors may impose a selective pressure on a pathogen which
can determine pathogen genotype and phenotype composition. Attenuation and evolution of myxoma
virus (MYXV) occurred when the virus was introduced and persisted in invasive European rabbits
(Oryctolagus cuniculus) in Australia and Europe, whereby the pathogen’s virulence decreased to
maximise transmission rates in field situations, alongside an increase in host immunity and survival
(Kerr 2012).
For optimal survival, a microbial species may be highly dependent on the availability of hosts and its
ability to exploit transmission opportunities between suitable hosts and adapt a generalist or specialist
4
lifestyle accordingly. This could be reflected in the microbial genome. High host specificity is thought
to limit gene flow within and between microbial species, and thus increase the genetic distance
between conspecific microbes infecting different hosts. This may explain the high levels of genomic
variation observed within some bacterial pathogens. For instance, 928 wild bird isolates of
Campylobacter jejuni were sequenced and compared to 1366 domestic animal and human isolates
(Griekspoor et al. 2013). Although there was a high level of diversity within the species, C. jejuni
sequence types clustered according to host, whereby C. jejuni genotypes from wild bird species were
different from each other and those from other sources (Griekspoor et al. 2013). Grouping in this
manner implies niche specialisation where there may be environmental and host associated barriers
restricting gene exchange (Griekspoor et al. 2013; Sheppard et al. 2011; Sousa & Hey 2013). In
contrast, a generalist lifestyle is thought to create more opportunities for gene exchange, and thus in
theory pathogens of the same species isolated from different hosts or locations should on average be
homogeneous. Supporting this hypothesis, high rates of gene flow were observed between the
generalist nematode Trichostrongylus axei infecting multiple sympatric ungulates in North America,
with most genetic variation structured within individual hosts (Archie & Ezenwa 2011).Four modes of
speciation are postulated to occur in nature based on barriers to gene flow from isolation, niche
formation and genetic polymorphism. These theories are termed: allopatric speciation, peripatric
speciation, parapatric speciation and sympatric speciation (Figure 1-1), and all are biologically
plausible within microbes with generalist or specialist lifestyles.
Figure 1-1: Theoretical hypotheses for the modes of speciation (adapted from http://en.wikipedia.org/wiki/Speciation)
5
1.1.3. The host population
Animal populations are often hierarchical with structural organisation into multi-level societies. On
the simplest level, individuals are connected to others through direct or indirect association to form
groups. Groups may be formed of familial aggregations or cohorts of solitary individuals sharing a
territory. Together conspecific groups form populations. Behaviour and lifestyle choices (Drewe et al.
2011), habitat availability (Almberg et al. 2012) and social structure (Caillaud et al. 2013; Nunn et al.
2014; Nunn et al. 2011) determine the organisation, density and frequency of contact within and
between individuals in a population which in turn may determine the likelihood of pathogen exposure
and transmission. Habitat influences on pathogen prevalence are also confounded by interactions
related to increased host density and connectivity within preferred sites (Almberg et al. 2012). For
example, grey wolves (Canis lupus) occupying prime habitat in Yellowstone National Park were most
susceptible to disease caused by the mite Sarcoptes scabiei (Almberg et al. 2012). Heterogeneities in
connectivity between susceptible and infected individuals within a population are key factors
determining the spread of infectious organisms. Not all infected hosts contribute equally to the
transmission of a pathogen in a population. The 20:80 rule is thought to apply in many populations,
whereby a few individuals contribute to most infections (Woolhouse et al. 1997; Woolhouse et al.
2005). These “super spreaders” play a central role in dissemination of a pathogen to many individuals
or species via high shedding and/or contacts rates (Lloyd-Smith et al. 2005; VanderWaal et al. 2013a).
The extent an individual is connected to others can contribute to its propensity to spread infection
(Christley et al. 2005) due to increased level of contact and opportunities to become infected.
Clustering of highly connected individuals within a population has been shown to contribute to the
growth rate of epidemics (Watts & Strogatz 1998).
Transmission dynamics of an infectious organism can vary temporally and geographically because of
differences in landscape and host attributes, and the ability of pathogens to persist in the environment
(Real & Biek 2007). Sharing of a pathogen may be determined by host behaviour resulting from
asynchronous space use, territoriality and shared resources (Nunn & Dokey 2006; Nunn et al. 2014;
Nunn et al. 2011). Faecal contamination of soil and water sources with environmentally stable
organisms such as Salmonella spp., allows indirect transmission of pathogens between independent
6
groups. Therefore, the spread of infectious agents between animals is not necessarily dependent on
coexistence. Vector borne pathogens spend a period of their lifecycle off-host, and thus are less
dependent on close proximity between hosts for persistence. For example, given suitable climatic
conditions the mosquito Culex quinquefasciatus is able to invade a new environment and disseminate
avian malaria (such as Plasmodium relictum) between multiple populations of susceptible birds, and
this has had a profound effect on Hawaiian honeycreeper populations in the Hawaiian archipelago
(Warner 1968).
1.1.4. Multi-host pathogens and reservoir dynamics
Many pathogens have complex lifestyles, colonising many species and environments. Multi-host
pathogens such as Salmonella spp., with over 2500 serotypes identified (Grimont & Weill 2007), are
able to colonise and infect a range of taxonomic groups. Commonly termed generalist organisms,
multi-host pathogens have significantly more complex transmission networks than that found in single
host pathogens. Each host species is effectively a sub-population within a larger framework of
susceptible hosts and different hosts and / or subpopulations may vary in their susceptibility and
immunity to infection. Heterogeneous contact patterns, behaviour and habitat preferences could all
influence colonisation and transmission of a pathogen within and between species sub-populations
(Dobson 2004).
Cross species transmission rates can be influenced by extent of genetic relatedness between hosts
(Huang et al. 2014). This is thought to be a key determinant of transmission and emergence of rabies
virus in different bat species in North America (Faria et al. 2013). Physiological and behavioural
similarity between taxonomically related hosts has been attributed to sharing of microbial subtypes of
Escherichia coli in wild ungulates in Africa (VanderWaal et al. 2014). However, even closely related
taxa can vary in their immunity to pathogens, where one host species may be more tolerant of a
pathogen than its competitor (Prenter et al. 2004). For example, the introduction of squirrel pox virus
to the red squirrel (Sciurus vulgaris) in Great Britain has contributed to the decline and ecological
replacement of the native squirrel with the closely related non-native North American grey squirrel
(Sciurus carolinensis), in which the virus is less pathogenic (Rushton et al. 2000; Sainsbury et al.
2008).
7
However, there does not need to be close taxonomic relatedness between the existing and new host for
a pathogen to infect a new host species. Chance transmission, including indirect contact between
transient individuals and resident populations can occur along wildlife corridors and might be
sufficient to cause a local epidemic (Hess 1994; Simberloff & Cox 1987). Use of corridors may be
biased towards one sector of the population, with those having larger ranges being prime vectors for
dissemination of disease between populations. Clements et al. (2011) investigated movements of
whitetail deer (Odocoileus virginianus) in the Midwest of America, in order to predict spread of
infectious diseases including chronic wasting disease. They found young males seeking new territory
dispersed at high rates along the river corridors, whilst others maintained high fidelity to their territory
(Clements et al. 2011). Therefore, young males seeking new territories were the most likely element
of the population to contribute to the dispersal of infectious diseases.
Multi-host pathogens are of concern to the conservation of endangered wildlife. Sympatric host
species harbouring unselective pathogens pose a disease threat for vulnerable species within
transmissible proximity. For example, Toxoplasma gondii type X may be contributing to the decline
of the threatened southern sea otter (Enhydra lutris nereis) in California. T. gondii type X has been
isolated and may be indirectly transmitted to the sea otters from infected wild and domestic felines in
the same region (VanWormer et al. 2014). Epidemiologically connected populations that act as
sources of pathogens to vulnerable populations are typically termed reservoirs (Haydon et al. 2002).
Reservoir hosts are classified according to their susceptibility to infection by an organism and ability
to maintain and transmit that organism without obvious detrimental effects (Johnson et al. 2012).
Despite extensive effort, determining the source of a pathogen and the directionality of transmission is
hindered by complex interactions and relationships which may be unobservable in wild populations.
These difficulties are apparent when trying to decipher the transmission of Mycobacterium bovis
between Eurasian badgers (Meles meles) and domestic cattle (Bos primigenius) (Biek et al. 2012).
Another example is brucellosis, which is a disease caused by the bacterium Brucella abortus.
Although it can threaten wild ungulate populations, much interest and conflict is created due to the
potential spread of B. abortus from reintroduced bison (Bison bison) to cohabiting domestic cattle in
North America. Although evidence suggests B. abortus may have originated from cattle introduced
into America, reviewed in Meagher and Meyer (1994), DNA typing of multiple isolates of B. arbortus
8
isolated from ungulates including wild elk (Cervus canadensis), bison, and cattle in the Greater
Yellowstone area revealed that elk and cattle shared similar bacterial genotypes, but bison isolates
were highly divergent from cattle sequence types (Beja-Pereira et al. 2009). This would imply that the
main transmission pathways may occur between elk and cattle.
Host switching and establishment of a pathogen in a new reservoir is not an unusual occurrence.
When the Australian brushtail possum (Trichosurus vulpecula) was introduced and established in
New Zealand (Pracy 1974), possum populations, became infected with Mycobacterium bovis in the
new location (Morris & Pfeiffer 1995). Interestingly, possums are free of Mycobacterium bovis in
their native range (Corner & Presidente 1981), Introduced possums have become a wildlife reservoir
for M. bovis, and are thought to be a ‘spill back’ source of infection to livestock (Morris & Pfeiffer
1995). However, the epidemiology of transmission is not well understood and as with many multi-
host pathogens in wildlife ecosystems, questions remain unanswered which hinder control of the
disease and the creation of management solutions. Investigations are exploring the intra- and
interspecies transmission of M. bovis through the use of epidemiological tools, including experimental
infection and social networking (Corner et al. 2003; Nugent et al. 2013; Rouco et al. 2013).
1.2. Conservation of wildlife in the face of disease threats
1.2.1. Disease threats posed to and from wildlife
Emerging infectious diseases (EIDs) in wildlife are of increasing concern due to their direct impact on
biodiversity and the risk of zoonotic transmission (Daszak et al. 2000; Jones et al. 2008). The
epidemiology of emerging infectious diseases (EIDs) in free living wildlife is complex, and our
understanding of the drivers of transmission and disease progression in these systems is hampered by
the fact that wild animals are difficult to observe and measure (McCallum et al. 2001). Three
hypotheses explaining the origins of wildlife EIDs are postulated: (i) spill-over of pathogens from
domestic animals, (ii) caused by anthropogenic drivers of infection, or (iii) with no involvement of
humans or livestock (Daszak et al. 2000).
Outbreaks of EIDs in nature have two major biological implications. First, wildlife EIDs can have
significant impacts on the biodiversity of free living animals, e.g. the role that the chytrid fungus
9
Batrachochytrium dendrobatidis has had in the decline of amphibian populations in Australia (Berger
et al. 1998). Second, the emergence of important infectious diseases in humans and livestock are often
reported to originating from wildlife reservoirs. For example, bats are frequently implicated as sources
of exotic human pathogens. Fruit bats (suborder Megachiroptera, family Pteropodidae of the order
Chiroptera) have been recognised as natural reservoirs of Hendra virus (HeV) in the Australasian
region in recent years (Plowright et al. 2011; Plowright et al. 2008). In this system, domestic horses
(Equus ferus caballus) are thought to be an intermediate host becoming infected through association
with fruit bats, with cases of HeV in humans an incidental occurrence after contact with an infected
horse (Plowright et al. 2011; Plowright et al. 2008). It is important to note that pathogen transmission
threats may not be mutually exclusive in one direction or the other. Bengis et al. (2002) highlight that
the interface between wildlife and livestock provides opportunities for bidirectional flow of
organisms, creating conflicts not only in domestic animals but also posing threats for co-habiting
wildlife.
In an increasingly connected world, humans and animals are frequently travelling between remote
locations with the potential to transmit exotic pathogens across barriers which were previously
unattainable. The introduced fungus Pseudogymnoascus destructans causes a disease termed white
nose syndrome in susceptible North American bats, with up to 95% mortality in some hibernacula and
killing over one million bats in North America (Frick et al. 2010). The fungus has been detected in
some bat species (Myotis myotis) in Europe (Pikula et al. 2012). However, it appears less pathogenic
in these populations and has not been associated with the mass mortalities observed on the American
continent (Cryan et al. 2013). It is thought that suitable environmental conditions, human travel and
the availability of susceptible naive hosts has allowed the rapid spread of the fungus upon introduction
to the eastern states of North America (Foley et al. 2011).
Stresses imposed by anthropogenic manipulation of the environment, habitat fragmentation,
introduction of invasive species, global travel and altered climate, often acting synergistically, have
contributed to the emergence of infectious disease and the declines of wildlife populations (Brook et
al. 2008; Munns 2006; Olival et al. 2013). Pathogenic outbreaks are rarely a sole driver of extinction
but in conjunction with additional selective pressures can pose a significant threat to a population
10
(Heard et al. 2013; Smith et al. 2006). A review of the IUCN red listed species revealed that less than
4% of extinctions and less than 8% of critically endangered species listed infectious disease as a
contributing factor to their decline (Smith et al. 2006). The threat of disease appears to increase as
population size decreases when a species, particularly amphibians and birds, moves further towards
extinction (Heard et al. 2013). However, these findings may be influenced by discovery and species
specific sampling bias whereby as a species heads towards extinction, the knowledge about threats
posed to that species increases (Heard et al. 2013). This would be an intuitive finding as wild animals
are notoriously difficult to observe in the wild, with diseased individuals often inconspicuous or
rapidly scavenged.
Fragmentation of habitats due to human activity has had negative impacts on wildlife populations
around the world (Foley et al. 2005; Plowright et al. 2008). For example, deforestation in the Peruvian
Amazon has altered breeding of a malarial mosquito (Anopheles darling) which has led to an
increased biting rate of humans in deforested areas (Vittor et al. 2005). Increased awareness of
isolated habitat patches and populations and the impact this has on the ecosystem has driven the need
for a solution. Habitat corridors are commonly proposed as a conservation tool to negate the effects of
fragmentation by facilitating plant and animal movement between isolated patches and increasing
population viability (Hilty et al. 2006). However, these corridors are nonspecific and have the
potential to negatively impact the same species that they are meant to benefit by inadvertently
facilitating the movement of unsolicited competitors, predators and the novel diseases they may carry
(Simberloff & Cox 1987). For example, Sullivan et al. (2011) found corridors had a positive effect on
spread of animal mediated transmission of plant parasites across highly connected landscapes.
Methods combining mathematical epidemiology and metapopulation modelling found that the spatial
arrangement and level of contact between populations within a metapopulation are important in terms
of pathogen spread, and that some arrangements can be better for disease control than others (Hess
1996).
Disease control to prevent zoonotic spread between the key players; humans, wildlife and livestock,
should be based on a broad understanding of the ecology and epidemiology of the disease agent
within its hosts (Woodford 2009). However, a challenge to studying transmission of disease including
11
wildlife is that the hosts are often inconspicuous and not easily observed, especially those with overt
disease.
1.2.2. Anthropogenic management of threatened wildlife populations: translocations and
sanctuaries as conservation tools
Translocation is the human mediated process of capture, movement and release of animals or living
organisms from one location to another (Soorae 2008). From this definition, translocation has been
occurring worldwide for centuries with the introduction of farming and trade of domestic and wild
animals. In more recent times, the intentional movement of individuals, populations and species
across landscapes is being used as a means to maintain biodiversity in the face of anthropogenic
changes to ecosystems (Sainsbury & Vaughan-Higgins 2012; Weeks et al. 2011). Translocation is
most often used and referred to in the context of the intensive management of threatened species,
where anthropogenic manipulation is required to increase or maintain a population (Weeks et al.
2011). In the United States, approximately 70% of all recovery plans for threatened and endangered
species have recommended this approach (Tear et al. 1993). The primary aim of a translocation for
conservation is survival and persistence of the species in the new location. Also translocation is being
proposed as a major tool for the conservation management of wild species unable to adapt to rapid
climate change (Mawdsley et al. 2009). The inadvertent introduction or emergence of infectious
disease through translocation of animals into new populations or ecosystems has become a major
concern when considering management of wildlife populations (Thompson et al. 2010).
Translocation of individuals between previously isolated ecosystems removes barriers to exchange of
pathogens (Power et al. 2013). These animals can potentially transfer exotic pathogens into extant
populations with no effective immunity at the release site (Anderson & May 1986; Woodford &
Rossiter 1994). Threats posed by management actions are particularly pertinent to endangered species
maintained in fragmented isolated populations and are heavily reliant on conservation measures for
the persistence of the species. Disease outbreaks can affect population dynamics and management,
with the extent of the impact often dependent on the size of the outbreak. Simulation of disease
epidemics in endangered Sierra Nevada bighorn sheep (Ovis canadensis) indicated severe outbreaks
of disease increased adult mortality, thus reducing population size and impeding population
12
management (Cahn et al. 2011). Although small epidemics did not impact population numbers, they
did reduce the number of individuals available for translocation, thus indirectly impacting population
management and viability (Cahn et al. 2011). Pathogenic incursions may not necessarily result in
mortality. Pathogen associated morbidity can impact hosts in subtle ways. For example animals
infected with Toxoplasma gondii show no overt symptoms of disease but the protozoan can influence
the behaviour of its host, increasing risk and likelihood of transmission between mice and cats
(Berdoy et al. 2000).
Although captive breeding and translocation of animals has become an important management tool
for threatened species programs worldwide (Fischer & Lindenmayer 2000; Griffith et al. 1989),
insufficient research has been conducted regarding the spread of disease associated with these
activities. The management of wildlife in artificial environments creates atypical opportunities for
exchange of microorganisms between humans, domestic animals and wildlife. During captivity
animals may be exposed to new infectious agents from previously unavailable transmission routes and
act as carriers and/or vectors when released into a new location (Viggers et al. 1993). Endemic island
animal species held in captivity can be particularly susceptible to infection by exotic pathogens. For
example, herpes virus infection was unknowingly transmitted from asymptomatic foster rock doves
(Columba livia) to captive-bred Mauritian pink pigeon chicks (Nesoenas mayeri) (Snyder et al. 1985).
Spill-over of human pathogens to animals (reverse zoonosis) in captivity is not well understood.
However, if a pathogen establishes in a wildlife reservoir, there is a potential for spillback to humans
(Thompson et al. 2010), or vice versa. A concerning example is that of reverse zoonotic pathogen
emergence in African apes. Increasing evidence indicates that human respiratory and gastrointestinal
pathogens are circulating in African gorillas and chimpanzees which are thought to have contributed
to significant morbidity and mortality in these populations (Palacios et al. 2011; Rwego et al. 2008;
Williams et al. 2008).
In theory, the population size of many endangered populations is thought to be below the threshold
required for establishment of disease (Mathews et al. 2006), particularly in small translocated sub
populations. However, reality is more complicated and alternative hosts in the form of sympatric
species within a given ecosystem may increase the effective population size and allow persistence of
13
an epidemic (Mathews et al. 2006), with devastating effects on the species of concern. Endangered
bighorn sheep (Ovis canadensis) in North America are have been translocated to re-establish or
augment existing populations (Boyce et al. 2011). Populations are small and are frequently afflicted
with epidemics of Mycoplasma ovipneumoniae associated with pneumonia and death (Besser et al.
2012a; Besser et al. 2012b). Pathogen spill-over from closely related domestic sheep (Ovis aries),
poly-microbial disease and the presence of carrier individuals are thought to contribute to localised
disease outbreaks in bighorn sheep, summarised in Besser et al. (2013). The presence of sympatric
domestic sheep may not only provide a source of infection, but could also effectively increase the
population size and in theory provide the suitable system required for maintenance of pathogens
within the threatened bighorn sheep populations.
Depending on the study species and ecosystem dynamics involved, translocated or captive reared
animals may be at an advantage in terms of host parasitism compared to their home range. The
parasite release hypothesis suggests that animals may escape native pathogens when moved to a new
location (Torchin et al. 2003). Many pathogens have complex life cycles requiring more than one
host, and thus if this host is not present in the new location, the pathogen may not be maintained. The
North American starling (Sturnus vulgaris) descended from a small population of introduced
European starlings (Sturnus vulgaris) in the 1800’s (Baker et al. 1986). Torchin et al (2003) report a
lower diversity of parasite species in the North American starlings compared to those of the source
population. The phenomenon may be explained by increased host resilience to parasites at the new
site (Torchin & Mitchell 2004), or alternatively host fitness may be higher due to lower levels of
parasite incidence and or exposure at the new location (Smith et al. 2009; Torchin & Mitchell 2004).
Translocated and reintroduced native species are regularly selected for good health prior to transfer,
consequently disease circulating in this select group may be lower than the source population
(Almberg et al. 2012). However, these populations are expected to acquire pathogens circulating
within the environment (Almberg et al. 2012), but this will often be at a slower rate than what they
would be exposed to in their original geographic range (Torchin et al. 2003; Torchin & Mitchell
2004). The dynamics of pathogen spill-over and spill-back have been summarised in a recent review
by Tompkins et al. (2011), using the example of the invasion of the North American signal crayfish
14
(Pacifastacus leniusculus) into the United Kingdom to demonstrate potential mechanisms of pathogen
transmission between introduced and native species (Figure 1-2).
Figure 1-2 Diagrammatic representation of pathogen transmission dynamics between native and introduced populations of crayfish in the United Kingdom (Tompkins et al. 2011)
1.3. Epidemiological tools for species conservation
1.3.1. Risk assessments and disease surveillance
Understanding the epidemiology of an infectious disease in free-ranging wild animals requires an
understanding both of the route of natural infection and of the processes underlying development of
clinical disease. A lack of knowledge of pathogen dynamics in wild animal populations limits the
ability of conservation managers to develop prioritised strategies for disease control and to effectively
target disease surveillance. It has been recognised that countries which invest efforts in wildlife
disease surveillance are more likely to be prepared for and respond to emerging infectious and
zoonotic diseases (Morner et al. 2002). The World Organisation for Animal Health (OIE) was
established in 1924 due to increased awareness of animal diseases (www.oie.int/about-us/history).
The OIE is an intergovernmental organisation with international policy objectives for the monitoring
15
and improvement of animal health worldwide (OIE 2011). Active disease surveillance in wildlife is a
relatively new concept and is fundamental for prediction and mitigation of disease epidemics in both
people and animals. Disease monitoring involves the “systematic recording of epidemiological data,
with the specific purpose of detecting spatial and temporal trends as well as presence or absence of the
disease” (Boadella et al. 2011). Pathogens selected for screening are often those that would have the
greatest impact on human health and the economy, or have a history of causing disease in threatened
wildlife populations (Boadella et al. 2011). Selective testing can create bias in detection, with a lean
towards exotic zoonotic diseases, especially those which are notifiable (Boadella et al. 2011; Hartley
& Gill 2010). Even when screening occurs, we are limited in our interpretation of results due to
imperfect tests resulting in false positive and negative results, as well as the intermittent shedding of
pathogens from hosts, as is often the case in faecal excretion of Salmonella spp. (Ivanek et al. 2012;
Van Immerseel et al. 2004). Detecting wildlife disease trends requires: knowledge of diagnostic test
sensitivity and specificity, adequate sample size over a significant time period, appreciation of the
logistics associated with sample collection, and expertise in data analysis where many parameters are
uncertain (e.g. total population size). Increased disease monitoring will provide detailed information
of pathogen diversity and relative importance in aetiology of disease in the host (Smith et al. 2009).
Improved baseline information with regards to the diversity and abundance of pathogens on a
community and ecosystem gradient will increase our grasp on potential transmission of pathogens
within and between wildlife and the threats posed to conservation (Thompson et al. 2010).
Globally, wildlife movements carry the risk of transfer of infectious agents and potential impact on
the health status of livestock, companion animals and humans (Cunningham 1996). Awareness of this
issue has prompted protocols to minimise disease transmission risk associated with wildlife
translocations and reintroductions (Cunningham 1996; Viggers et al. 1993; Woodford & Rossiter
1993). The OIE use an assessment model to analyse the disease risk associated with movements of
animals (Murray 2004). Several models have attempted to estimate the likelihood that each pathogen
would cause disease by assessing the probability and consequences of release and exposure in a new
location (Armstrong et al. 2007; Sainsbury & Vaughan-Higgins 2012). Models classified risk ranging
from negligible to high and evaluated management options according to outcome. Sainsbury and
Vaughan-Higgins (2012) attempted to assess the disease risk of translocating wild animals using the
16
relocation of the Eurasian crane (Grus grus) from Germany to England as a proxy for other species.
However, they acknowledge there are limitations in interpreting the outcome of models due to
incomplete identification of pathogens affecting the target species. Therefore, animals may be
translocated with unknown pathogens which could impact health and population reestablishment
(Ewen et al. 2012; Sainsbury & Vaughan-Higgins 2012). For example, three endangered juvenile
kakapo (Strigops habroptilus) died shortly after translocation to a new location in New Zealand due to
an outbreak of the bacteria Erysipelothrix rhusiopathiae, which may have been contracted from soil at
some stage of transport or arrival, or manifested as a result of the stress of translocation (Gartrell et al.
2005). These animals had been screened for diseases of concern prior to translocation but erysipelas
had not been considered as an issue for the parrots prior to this outbreak (Gartrell et al. 2005). This
highlights one of the problems in predicting EIDs in small wildlife populations.
Encounters with new pathogens may occur at several stages of the process, including before, during
and after translocation (Cunningham 1996). Despite the risk of pathogen emergence and spread, many
wildlife relocations occur without disease screening. A report found less than a quarter (24%) of
translocations which occurred in Australia, Canada, the United States and New Zealand used a health
screening protocol prior to movement of animals (Griffith et al. 1993). The primary aim of disease
screening associated with translocations is to prevent introduction of novel pathogens into existing
populations at the release site, as well as increasing the likelihood of survival of translocated
individuals (Cunningham 1996; Viggers et al. 1993; Woodford & Rossiter 1993). There are few laws
and regulations in place to enforce biosecurity and disease screening and as previously mentioned
there is substantial variability in uptake of disease assessment associated with translocations (Daszak
et al. 2000). Nonetheless, conservation interest in quantifying the likelihood of disease transmission
has increased reports of health screening, including a recent health assessments of translocated
western ring tailed possums (Pseudocherirus occidentalis) (Clarke et al. 2013), water voles (Arvicola
terrestris) in the United Kingdom and marsupial dibblers (Parantechinus apicalis) in Australia
(Mathews et al. 2006).
17
1.3.2. Pathogen ecology and epidemiology
Molecular characterisation of pathogens derived from multiple sources provides insights into the
epidemiology of infectious disease ecology. When combined with epidemiological data, genetic tools
can be used to identify hosts, investigate pathogen adaptation, infer chains of transmission and link
heterogeneities in pathogen prevalence to host and environment associated factors (Archie et al.
2009). An epidemiological investigation of Mycobacterium tuberculosis isolates from humans in an
outbreak in a community in British Columbia, Canada, nicely demonstrates the recent progression of
epidemiological modelling in an applied context (Gardy et al. 2011). In this example, researchers
integrated social network contact data derived from questionnaires with whole genome sequencing of
M. tuberculosis isolates from patients. Sequence analysis revealed two genetically distinct lineages of
M. tuberculosis circulating in the community, indicating that there were concomitant outbreaks; an
observation that was not previously readily intuitive from standard molecular methods (Gardy et al.
2011). Additionally, epidemiological investigation identified that the outbreak coincided with
increased crack cocaine use in the community and social connections (Gardy et al. 2011). If
conducted early in a disease outbreak, integrative studies of this manner may be used to prevent or
reduce transmission routes and thus temper the extent of an epidemic.
Fine scale microbial genetics can be used to infer who infected who in a transmission chain based on
sequence profiling. Pathogen subtypes are assigned according to genetic similarity or differences.
Therefore, if two individuals share a subtype, transmission may be inferred (Bull et al. 2012;
VanderWaal et al. 2013b, 2014). However, limitations are still evident in this technique as direction of
transmission cannot be determined without further investigation. Additionally, sharing of a subtype
may be indicative of exposure to a common source or similar life history rather than direct
transmission between two hosts.
Population genomics has been used in order to attribute niche adaptation of the pathogen to the host or
environment. A study of Campylobacter jejuni in New Zealand used multi locus sequence typing
(MLST) of bacteria isolated from humans and attributed infections to the consumption of chicken
from certain poultry suppliers due to the genomic relatedness between human and chicken associated
isolates (Mullner et al. 2010). However, comparison of a few genes may not capture the full genomic
18
diversity present between closely related species. Ribosomal multi locus sequence typing (rMLST) of
53 conserved genes encoding ribosomal proteins on the bacterial genome is an alternate method for
more defined sequence type comparisons (Jolley et al. 2012). It has been used for taxonomic
exploration of Neisseria species, revealing previously unobserved complexity within Neisseria
polysaccharea suggesting it could comprise more than one taxonomically distinct organism (Bennett
et al. 2012).
In nature, some individuals interact more than others and thus the likelihood of transmission and
infection with a pathogen is likely to follow suit. Social network analysis attempts to mathematically
quantify the variation in connectivity and pathogen transmission pathways in relation to host or
population biotic and abiotic factors (Newman 2010). For example, in a transmission study
investigating Mycobacterium bovis infection in captive brushtail possums (Trichosurus vulpecula),
infected possums had higher connectivity measures than those which were uninfected (Corner et al.
2003). Network concepts can be applied to identify targets to mitigate pathogen spread and thus are a
useful tool to study the demographics of infectious diseases in wildlife (Godfrey 2013; Rushmore et
al. 2013; VanderWaal et al. 2013a; VanderWaal et al. 2013b, 2014).
1.4. Research focus
1.4.1. New Zealand conservation management
Worldwide, island ecosystems are particularly vulnerable to extinctions. In New Zealand, 41% of all
native bird species have become extinct since human settlement (Tennyson & Martinson 2006).
Invasive non-native mammalian predators and anthropogenic activities have a long history of impact
on the populations of New Zealand’s native flora and fauna. As a result, New Zealand has been
innovative with its conservation management of endangered species. To prevent further extinction of
land birds, translocation and mammal eradication emerged as conservation solutions and as a result
numerous offshore islands were utilised for reasons of isolation and conservation value (Sherley et al.
2010). The process of removing exotic animals from island reserves began in the early 20th Century
and since then the islands have provided sanctuary for many indigenous populations threatened by
introduced mammals (Bellingham et al. 2010). Conditions created due to human disturbance have
19
allowed non-native birds to colonise these islands, with starlings (Sturnus vulgaris) observed as far as
Campbell Island in the Sub-Antarctic (Heather & Robertson 2006). Introduced avifauna may compete
with natives who occupy the same ecological niche (Bellingham et al. 2010) and may act as reservoirs
of disease (Rushton et al. 2006).
Translocations have become a common conservation management strategy for numerous endangered
birds and reptiles in New Zealand and many of these have been catalogued (Sherley et al. 2010). A
particularly successful example is the North Island saddleback (Philesturnus rufusater) which has
been subject to multiple translocations and has successfully increased its range from a single source
population to occupying at least 13 islands (Parker 2008). However, not all translocations have been
successful. The translocation of hihi (Notiomystis cincta) to offshore islands has been attempted but
has not been effective with remaining populations heavily reliant on supplementary feeding for
maintenance (Armstrong et al. 2007).
Translocations of animals between reserves are actively managed by the New Zealand Department of
Conservation (DOC), or by local conservancies and communities under the guidance and control of
DOC. All known translocations in New Zealand are performed in accordance with the DOC disease
risk assessment tool, a flow diagram that assesses risk associated with infectious disease in the source
population as well as the target population (McInnes et al. 2004). With all good intentions, risk
assessments may not be able to predict outbreaks of some diseases, with some translocation associated
deaths reported (Alley et al. 2008; McLelland et al. 2011).
20
1.4.2. Takahe (Porphyrio hochstetteri)
Figure 1-3 Adult takahe (Porphyrio hochstetteri) resident on Maud Island, New Zealand. Photo courtesy of Thomas Burns
The takahe (Porphyrio hochstetteri) (Figure 1-3) is an iconic endemic New Zealand flightless rail,
considered “nationally critical” under the New Zealand DOC threat classification system (Miskelly et
al. 2008) and “endangered” on the IUCN Red list of threatened species (BirdLife International 2013).
Takahe are a species of high cultural value and are regarded as a “taonga (treasured) species” to the
Maori people of the South Island, Ngai Tahu. As a result, takahe are under strict protection to
conserve the species, with consultation required when decisions are made with regards to their
welfare. The problems facing takahe are representative of most of New Zealand’s endangered bird
species. The modern day takahe were thought to have been widespread throughout the South Island,
New Zealand, while a closely related extinct species of takahe (Porphyrio mantelli) inhabited the
North Island (Trewick 1996; Trewick & Worthy 2001). All takahe were thought extinct by the end of
the 19th Century until a small population of approximately 250 South Island takahe were rediscovered
in Fiordland, South Island in 1948 (Ballance 2001). It is believed that their range was reduced to
Fiordland as a consequence of confounding attributes including: hunting by Maori, habitat destruction
and the introduction of mammalian predators and competitors (Bunin & Jamieson 1995; Trewick &
Worthy 2001).
21
Over the next 15 years, a 518-km2 takahe conservation area was created in the Murchison mountains,
Fiordland (Ballance 2001), and subsequently takahe from the Fiordland population were introduced to
five small predator free offshore islands (Crouchley 1994). Takahe were first translocated to Maud
Island in 1984, followed by Mana Island in 1988, Kapiti Island in 1989 and Tiritiri Matangi Island in
1991, with additional introductions in subsequent years (Bunin et al. 1997). Takahe are currently
located in 16 reserves or sanctuaries across New Zealand (Figure 1-4). Current management efforts
are focused towards predator trapping, nest manipulation, captive rearing and regular relocation of
birds between sanctuaries (Hegg et al. 2012).
The takahe population has gone through events of rise and decline in recent times, with a current
approximate population of 227 adult birds, all of which are located in New Zealand (Figure 1-4)
(Wickes et al. 2009). There still remains a free living population of takahe in the Murchison
Mountains, Fiordland, as well as populations in a dedicated breeding facility and on offshore and
mainland sanctuaries (Figure 1-4) (Wickes et al. 2009). The conservation aim in establishing these
sub-populations was to provide insurance populations and mitigate against both deterministic and
stochastic threats (Lee 2001). All the birds external to the Murchison Mountains are descendants of
the small original Fiordland population and as such are managed as a single population with multiple
translocations to minimise the potential detrimental effects of inbreeding (Jamieson et al. 2006). Even
so, island populations originate from a small population and this coupled with a low carrying capacity
has led to inbreeding and loss of genetic diversity (Bunin et al. 1997; Jamieson & Ryan 2001). Island
populations have been less productive and have shown elevated egg infertility compared with the
Fiordland and captive breeding populations (Grueber et al. 2010). In addition, inbreeding depression
has been linked to in increased susceptibility to disease, due to reduced variation of gene loci in the
major histocompatibilty complex resulting from lack of external genetic input into the population
(Agudo et al. 2012; Obrien & Evermann 1988). Impaired ability to elicit an immune response and
susceptibility to infection is a distinct possibility in such a small population of takahe. However,
disease has rarely been questioned as a contributing factor in the decline of the takahe population. A
retrospective investigation of post mortem takahe records revealed that 20% of adult deaths were
attributable to infectious or inflammatory disease, with the infectious organisms Erysipelothrix
rhusipathiae and Salmonella enterica serotype Typhimurium cultured from dead takahe (McLelland
22
et al. 2011). This study noted the lack of cohesive records of disease within this endangered species
(McLelland et al. 2011). In the early days of takahe recovery management, cross-fostering was used
where a pukeko (Porphyrio porphyrio) foster mother was trained to rear foster takahe chicks (Bunin
& Jamieson 1996). Although the two species are known to share some habitat and are closely related,
this practice may have unwittingly introduced exotic pathogens to the takahe population.
A health assessment of the takahe population highlighted that some infectious and parasitic agents
may have been introduced to takahe from activities associated with their management (Rose 2000).
However, there were no significant findings and it was recommended that a thorough quarantine
protocol should be applied to any movements between takahe populations (Rose 2000). Each takahe
movement between locations now requires the completion of DOC regulated disease screening and
quarantine/biosecurity protocols (McInnes et al. 2004). Active management by DOC involving
translocations, conservation strategies and on-going research investigating the health and welfare of
takahe is now required to conserve the species (Wickes et al. 2009).
Figure 1-4 Map of the current distribution of takahe in New Zealand. A private island location used as a breeding reserve is undisclosed due to confidentiality restrictions
23
1.5. Microbes of interest to this study
1.5.1. Campylobacter species
Campylobacter spp. have been isolated from humans (Kittl et al. 2013), domestic animals (Colles et
al. 2011; Kittl et al. 2013; Mughini Gras et al. 2012) and wildlife, in particular birds (Colles et al.
2011; Colles et al. 2009; French et al. 2009; Hughes et al. 2009; Keller et al. 2011; Sippy et al. 2012)
throughout the world. The organisms can survive in host faecal matter and can contaminate the
environment including rivers, streams and lakes (Jones 2001). Therefore, Campylobacter spp. can
transmit non-selectively between different hosts via indirect routes (Kwan et al. 2008). However,
Campylobacter spp. genotypes demonstrate a degree of host association (Colles et al. 2009;
Griekspoor et al. 2013; Mohan et al. 2013). Source attribution studies have extensively investigated
the transmission and carriage of the bacteria within and between host species (Colles et al. 2011;
French et al. 2009; Sheppard et al. 2010). Observation of the extent of genomic diversity in
Campylobacter spp. has been made possible by the use of genomic typing primarily by multi-locus
sequence typing (MLST) (Maiden et al. 1998). Many studies have used these genomic profiles for
sequence comparisons between Campylobacter spp. isolates from multiple sources (Colles et al. 2011;
French et al. 2009; Hughes et al. 2009; Kittl et al. 2013; Mohan et al. 2013) and have provided insight
into the epidemiology and evolution of Campylobacter spp. in multiple hosts.
Campylobacter spp. of relevance to animal and human health including Campylobacter jejuni and
Campylobacter coli are frequently isolated from faecal material derived from avian populations
(Waldenstrom & Griekspoor 2014) exhibiting no obvious clinical signs of disease (Benskin et al.
2009). Therefore, it is thought that Campylobacter spp. are a common commensal component of the
avian intestinal microflora. However, there are occasional reports of pathogenic effects in birds
(Waldenstrom et al. 2010). The prevalence of Campylobacter spp. in birds vary widely between taxa
and ecological guilds (Waldenstrom & Griekspoor 2014), ranging from 25% carriage in gulls (Keller
et al. 2011) to 78% in coots (Antilles et al. 2013). However, many of these studies have focused on
Campylobacter spp. of significance to human health and of socio-economic importance, omitting
potential incisive information available in unidentified Campylobacter spp.
24
A diverse range of Campylobacter spp. genotypes have been reported in domestic and wild birds
which are largely host specific and have been associated with ecological attributes relating to
domestication, habitat and behaviour (Benskin et al. 2009; Colles et al. 2011; Mohan et al. 2013;
Waldenstrom et al. 2002). Host associated factors are thought to be more influential on sequence type
carriage than geographical separation, with similar sequence types found in the same avian species on
different continents (French et al. 2009; Sheppard et al. 2010). Host behaviour and the ability of birds
to fly and migrate between locations may explain the mixing of genotypes between biogeographic
regions, with limited gene flow between species due to barriers related to occupation of distinct
ecological niches.
1.5.2. Salmonella species
Salmonella spp. are ubiquitous opportunistic bacteria and have been isolated from multiple wildlife
species around the world (Marin et al. 2013; Marin et al. 2014; Middleton et al. 2014; Vlahović et al.
2004). These organisms can survive outside their hosts for extended periods of time, are prime
candidates for contamination of the environment (Winfield & Groisman 2003), and show high
potential for cross-infection between susceptible hosts. Avian populations can be asymptomatic
carriers of Salmonella spp. (Vlahović et al. 2004). Even so, Salmonella spp. can cause disease in bird
populations and has been a great contributor to passerine mortality in the United States (Saito & Hall
2008). Salmonella enterica subspecies enterica serotype Typhimurium has been implicated as a cause
of death in both native and introduced birds in New Zealand, with a previously undetected strain in a
native range restricted passerine (hihi, Notiomystis cincta) located on a remote island in New Zealand
(Alley et al. 2002; Ewen et al. 2007). Ubiquity of Salmonella spp. in the environment and wildlife
reservoirs, for example, sparrows (Alley et al. 2002) and reptiles (Middleton et al. 2014), brings cause
for concern when considering the management of vulnerable wildlife species. Disease screening
protocols associated with translocations in New Zealand include testing for Salmonella spp. carriage
(McInnes et al. 2004), with the aim of selecting pathogen free individuals and prevent the spread of
novel Salmonella spp. to naive populations.
25
1.6. Objectives of the study
The overarching hypothesis of this study is that translocation, population fragmentation and isolation
have altered host-microbial dynamics in endangered takahe.
There are four main objectives of this study. The first is to examine population connectivity between
takahe subpopulations with respect to pathogen dynamics and identifying targets for surveillance
(Chapter 2). The second is to determine the prevalence of a common commensal, Campylobacter spp.
(Chapter 3) and an opportunistic pathogen, Salmonella spp. (Chapter 6) in multiple populations of
biogeographically separated takahe. The third is to investigate the influence of geographic isolation
and translocation on the molecular differentiation of Campylobacter spp. carried by takahe (Chapter
4). The final objective is to determine reservoirs and transmission dynamics of Campylobacter spp. in
vertebrate communities within an island ecosystem used for conservation of takahe (Chapter 5).
26
CHAPTER 2
NETWORK ANALYSIS OF TRANSLOCATED TAKAHE
POPULATIONS TO IDENTIFY DISEASE SURVEILLANCE TARGETS
Zoë L. Grange, Mary van Andel, Nigel P. French and Brett D. Gartrell
Conservation Biology, Volume 28, No. 2, 518-528
27
28
2. Network analysis of translocated takahe populations to identify disease
surveillance targets
2.1. Abstract
Social network analysis is increasingly being used in epidemiology and disease modelling in humans,
domestic animals, and wildlife. This study investigated a translocation network (area that allows
movement of animals between geographically isolated locations) used for the conservation of an
endangered flightless rail, the takahe (Porphyrio hochstetteri). Records of takahe translocations within
New Zealand were collated, and we used social network principles to describe the connectivity of the
translocation network. That is, networks were constructed and analysed using adjacency matrices with
values based on the tie weights between nodes. Five annual network matrices were created using the
takahe data set, each incremental year included records of previous years. Weights of movements
between connected locations were assigned by the number of takahe moved. The number of nodes
(itotal) and the number of ties (ttotal) between the nodes were calculated. To quantify the small-world
character of the networks, the real networks were compared to random graphs of the equivalent size,
weighting, and node strength. Descriptive analysis of cumulative annual takahe movement networks
involved determination of node-level characteristics, including centrality descriptors of relevance to
disease modelling such as weighted measures of in degree (kiin), out degree (ki
out), and betweenness
(Bi). Key players were assigned according to the highest node measure of kiin, ki
out and Bi per network.
Networks increased in size throughout the time frame considered. The network had some degree of
small-world characteristics. Nodes with the highest cumulative tie weights connecting them to other
nodes were the captive breeding centre, the Murchison Mountains and two offshore islands. The key
player fluctuated between the captive breeding centre and the Murchison Mountains. The cumulative
networks identified the captive breeding centre every year as the hub of the network until the final
network in 2011. Likewise, the wild Murchison Mountain population was consistently the sink of the
network. Other nodes, such as the offshore islands and the wildlife hospital, varied in importance over
time. Common network descriptors and measures of centrality identified key locations for targeting
disease surveillance. This technique provides a visual representation of animal movements in a
population that can aid decision makers when they evaluate translocation proposals or attempt to
control a disease outbreak.
29
2.2. Introduction
The introduction of a an infectious disease into small, highly connected wildlife populations can have
severe consequences on survival of the population (Daszak et al. 2000). Understanding how a
pathogen may spread from the point of entry throughout a connected community can inform
important transmission pathways, and thus aid in planning disease surveillance, outbreak control, and
contingency planning (Christley & French 2003). Dissemination of an infectious agent within a
population is dependent on many factors, including but not limited to susceptibility, mode of
transmission, and social interaction (Wey et al. 2008). Analysis of social relations of individuals and
identification of those that are more social within a group might identify individuals that are most
influential in pathogen transfer. Social network analysis (SNA) allows the pattern of contacts within a
complex system to be described and quantified in terms of a network of links between items. SNA is
becoming an important tool in epidemiology of human populations. Thus, its use in wildlife settings is
a natural progression.
SNA highlights which objects and relations are most important for maintaining connectivity of the
network, leading to its use in epidemiological disease modelling. Studies have been conducted
describing contact networks in relation to the transmission of infectious agents in animals (Christley
& French 2003). However, there have been no reported applications of network analysis to wildlife
translocation databases of endangered species to identify high-risk locations for the prioritization of
target sites for disease surveillance.
Infectious diseases are increasingly recognized as a challenge to the conservation of wildlife,
particularly in intensively managed species (Smith et al. 2009). A lack of knowledge of pathogen
dynamics in wild animal populations limits the ability of conservation managers to develop prioritised
strategies for disease control and to effectively target disease surveillance. When animals are
translocated their parasites and pathogens can also move with them (Bengis et al. 2002). Therefore
there is a need for measures to be in place to reduce disease risk during a translocation program
(Viggers et al. 1993). The primary aim of disease screening associated with translocations is to
prevent introduction of novel pathogens into existing populations at the release site and to increase
likelihood of survival of translocated individuals (Viggers et al. 1993).
30
Island ecosystems worldwide are particularly vulnerable to extinctions. In New Zealand, 41% of all
bird species have become extinct since human settlement and the introduction of invasive species
(Tennyson & Martinson 2006). To prevent further extinction of land birds, translocation and mammal
eradication emerged as conservation solutions. The process of removing exotic animals from island
reserves began in the early 20th century, and the islands have provided sanctuary for many indigenous
populations threatened by introduced mammals (Bellingham et al. 2010). Translocations of animals
between these sites are commonplace and are actively managed by the Department of Conservation
(DOC) or by local conservancies and communities under the guidance and control of DOC (Sherley et
al. 2010). However, not all translocations have been successful, for reasons often unknown or
unexplored.
The takahe (Porphyrio hochstetteri) is an iconic endemic New Zealand flightless rail, considered
nationally critically endangered under the New Zealand Department of Conservation threat
classification system (Miskelly et al. 2008) and endangered by the International Union for
Conservation of Nature (BirdLife International 2013). Takahe are thought to have been widespread
throughout New Zealand. Their range was reduced to Fiordland, South Island, as a consequence of
hunting by Maori and habitat destruction (Trewick & Worthy 2001). The species was thought to be
extinct by the end of the 19th century, but a small population was rediscovered in Fiordland in 1948
(Ballance 2001). Over the next 15 years, takahe were introduced to five small predator free offshore
islands to provide insurance populations and mitigate against both deterministic and stochastic threats
(Lee 2001). The takahe population has fluctuated in recent times, with a current approximate
population of 227 adult birds, all of which are located in New Zealand (Wickes et al. 2009). There
remains a free-living population of takahe in the Murchison Mountains, Fiordland, and insurance
populations on predator free offshore island reserves, mainland sanctuaries, and a dedicated breeding
facility (Wickes et al. 2009) (Appendix 9.2-1). To conserve the species, management strategies have
been targeted toward captive breeding and annual translocations of the birds between reserves
(Wickes et al. 2009).
This study investigated the use of social network theory and measures of node-level dynamics to
describe connectivity patterns and identify surveillance targets in isolated populations of translocated
31
takahe. Applying comparative analysis to random networks, this study determined whether the takahe
networks possess small-world properties. That is, a network in which most nodes are not neighbours
of one another but can be reached from every other by a small number of steps. Small-world
properties are a measure of particular relevance to population disease spread dynamics. This
investigation examined annual cumulative yearly changes in the translocation network, identified key
player (structurally important) locations in each network, and determined the importance of these for
maintaining the integrity of the system and thus their role in pathogen spread.
2.3. Methods
2.3.1. Data set
A census of all recorded takahe movements within New Zealand from 2007 to 2011 was collated into
a database. Prior to 2007, accurate records of takahe movements were not maintained. Records
analysed in this investigation include a DOC document of 131 takahe translocations within New
Zealand between a breeding centre, island reserves, and mainland reserves (Appendix 9.2-2). In
addition, 33 catalogued records (patient origin and release location) of hospitalized takahe from the
clinic database at Wildbase Hospital, Massey University were included in the analysis. In the context
of this analysis, the term takahe refers to takahe eggs, chicks, and adults. Although life stages may
carry variable risk of disease transmission, life stages were not separated due to uncertainty of the data
provided with regards to age of individuals. One database (Appendix 9.2-2) was analysed in two
formats for subsequent network analysis. The dataset consists of the year of the movement, an origin
and destination location, and number of takahe moved between the locations. Due to scarcity of
information with regards to precise dates of animal movements, each individual takahe moved was
considered as a single event. Therefore, within the same year, two birds translocated at the same time
would be considered the same as two birds translocated months apart. Both situations would be
regarded as two movement events and weighted accordingly. Number of contacts and direction of
movement between locations was known, and in the context of this study, the dataset has been
interpreted in both the directed and undirected format. For ease of understanding, each location has
been assigned a letter (Figure 2-1), and for subsequent analysis locations are referred to in this format.
32
For the purposes of this analysis, a network is referred to as the entirety of the interactions for a given
data set. A node is the location where takahe have been moved to and from. Ties represent the
connection, unidirectional, or bidirectional, between locations due to movement of birds. Tie weights
were derived from the number of takahe moved between 2 given nodes. Node strength was calculated
as the total tie weights associated with a node.
2.3.2. Network description and topology
Networks were constructed and analysed using adjacency matrices with a range of values based on the
tie weights between nodes. Five annual network matrices were created using the takahe data set, each
incremental year included records of previous years. For example, matrices were composed for 2007,
2007–2008, 2007–2009, and so on until 2011. Therefore, annual networks account for potential
reconnection of previously isolated nodes into the network. Weights of movements between
connected locations were assigned by the number of takahe moved within the year or cumulative
years included in the network. Takahe network matrices were analysed for weighted network
measures using tnet (Opsahl 2009) and igraph (Csardi & Nepusz 2006) packages within R software (R
Core Team 2013).
Each network was described in terms of size. The number of nodes (itotal) and the number of ties (ttotal)
between the nodes were calculated. These parameters allowed comparison between years and
monitoring of growth or reduction in the translocation network over time.
Understanding network topology and the pattern of contacts within a population can help predict
disease transmission pathways (Christley & French 2003). Small-world networks, as opposed to
regular and random networks, can be described as having clusters of connected nodes that have short
paths to connect nearby and distant groups within the network (Watts & Strogatz 1998). Small-world
determination is of relevance to disease simulations due to the influence such a connected system
could have on the rapid spread of disease from local infections across apparent long distances (Watts
& Strogatz 1998). For the purposes of evaluation of small-world characteristics, undirected weighted
annual matrices were created from the takahe dataset. Cohesion characteristics including weighted
geometric mean clustering coefficients (CCav) (Opsahl & Panzarasa 2009) and weighted average
33
shortest path lengths (Lav) (Opsahl et al. 2010) were calculated for each takahe network using
appropriate measures previously described within the tnet package. The clustering coefficient
measures the average proportion of connections that exist between locations divided by the number of
possible connections that could have existed (Dube et al. 2009). This measure was used to compare
the structure of the annual networks.
For all networks, the tnet package was used to create a random graph of the equivalent size,
weighting, and node strength; weights were locally reshuffled on each node across its outgoing ties
(Opsahl et al. 2008). Mean weighted measures of CCav and Lav were calculated from the random
network per annual observed network (Opsahl et al. 2010; Opsahl & Panzarasa 2009). Random graph
measures were then compared with the mean CCav and Lav of the observed annual takahe networks, in
order to determine whether each network has small-world properties (Watts & Strogatz 1998). To
quantify the small-world character of the networks and thus allow comparison over time, we divided
the mean Lav by the CCav for both the random and real networks (De Nooy et al. 2011).
2.3.3. Network dynamics
Descriptive analysis of the cumulative annual takahe movement networks involved determination of
node-level characteristics, including centrality descriptors of relevance to disease modelling such as
weighted measures of in degree (kiin), out degree (ki
out), and betweenness (Bi) (defined below). Key
players were assigned according to the highest node measure of kiin, ki
out, and Bi per network
(Appendix 9.2-3).
Weighted betweenness (Bi) is a measure of centrality that represents the likelihood of a node
connecting two random nodes via the shortest path accounting for node strength. A node's strength in
a weighted network is given by the sum of the weights of its ties. Locations with the highest weighted
betweenness measure are most likely to be involved in funnelling an infectious disease among other
nodes through the network via the shortest path.
Weighted degree (ki) is a complimentary centrality measure to betweenness, values of which are
attributed according to a tuning parameter that accounts for the traditional measures of number of
direct ties associated with a node, and the average number of tie weights of that node (Opsahl et al.
34
2010). In other words, weighted degree measures the number and strength of direct transfers between
a location and its neighbour. In a disease simulation network, weighted degree measures are positively
correlated with probability of becoming infected or transmitting infection and therefore could identify
locations for disease surveillance. The ability to calculate and compare in degree (kiin) and out degree
(kiout) allows identification of nodes as hubs or nodes that are centrally placed due to high numbers of
inward or outward connections to their neighbours. Additionally, from a disease surveillance
perspective, nodes on the periphery of a network could be just as important as those that are centrally
placed. Some nodes could act as end points for an infectious disease (sinks) or can harbour novel
pathogens that have not yet infiltrated a network (source). Nodes with high kiin and low ki
out scores are
a good indication of potential sinks and or sources.
2.4. Results
2.4.1. Network description and topology
Networks increased in size throughout the timeframe included in this analysis (Table 2-1). This was
demonstrated by the incremental number of nodes and ties included in the network each year. These
measures suggest that the annual networks were composed of several clusters of well-connected nodes
with high density and weight of ties. Over the years the Lav of the observed networks remain
consistently short, between three and four nodes, similar to that of the random graphs, a characteristic
of small-world networks. However, in the comparison of the Lav/CCav of takahe networks to that of the
random networks, only the 2007 to 2010 network showed some degree of small-world characteristics,
represented by a lower Lav/CCav than that of the random networks.
2.4.2. Network dynamics and node-level analysis
Cumulatively nodes with the highest tie weights connecting them were nodes A (captive breeding
centre) and Q (wild Murchison Mountain population), A and C (Mana Island), and A and J (Tiritiri
Matangi Island) (Figure 2-1).
35
Average path length (Lav)b Clustering coefficient (CCav)c Small world measure (Lav/CCav)d Year Nodes (itotal ) Ties (ttotal) Observed Random Observed Random Observed Random 2007 6 9 3 3 0.3 0.4 10 9 2007 to 2008 8 20 3 2 0.7 0.5 4 5 2007 to 2009 13 30 4 4 0.6 0.5 6 7 2007 to 2010 14 34 4 5 0.5 0.4 8 12 2007 to 2011 17 43 4 4 0.5 0.4 8 9
aThe takahe translocation networks represent the movement of takahe between locations used for their conservation management within New Zealand. Weights of movements between connected locations were assigned by the number of takahe moved within the year. The number of nodes (itotal) and the number of ties (ttotal) between the nodes were calculated for each network. To quantify the small-world measure of the networks, the real networks were compared to random graphs of the equivalent size, weighting, and node strength. bAverage path length (Lav) is the number of steps required to move between two randomly assigned locations in the network. cClustering coefficient (CCav) is a measure of degree to which nodes in the network tend to cluster together. dSmall-world measure is the degree to which most nodes in the network are not neighbours of one another but can be reached from every other node by a small number of steps.
Table 2-1 Measures of observed weighted takahe translocation networks and corresponding random measures which were then compared with the observed annual takahe networks to determine whether each network has small-world propertiesa
36
Figure 2-1 Node (circles) and tie (lines) networks of cumulative takahe translocations from 2007 to 2011: (a) tie weight between locations (the thicker the line, the more translocations between two directly connected locations) and (b) weighted betweenness for individual locations (i.e., measure of centrality determining the likelihood of a node connecting two random nodes via the shortest path while accounting for node strength [sum of node weights] of an individual node). Nodes represent isolated geographic locations used for takahe conservation. Lines with either uni- or bidirectional arrows show the direction of human-mediated takahe translocations between locations. Size of the circle illustrates extent of centrality; the larger the node, the more central the location in the network (A, Burwood Bush breeding centre; B, Kapiti Island; C, Mana Island; D, Maud Island; E, Maungatautari reserve; F, Pukaha Mt Bruce; G, Wildbase Hospital; H, Private island; I, Te Anau wildlife reserve; J, Tiritiri Matangi Island; K, Motutapu Island; L, Peacock Springs wildlife park; M, Secretary Island; N, Wellington Zoo; O, Willowbank reserve; P, Zealandia / Karori Sanctuary; Q, Murchison Mountains).
37
Identification of key players according to the highest measure of weighted degree (ki) when
accounting for historic connections demonstrated relatively consistent key players over time (Table
2-2). The key player according to in degree (kiin) fluctuated between a captive breeding centre (node
A) and the wild Murchison Mountains population (node Q). Node Q had a large number of takahe
inputs in 2008 and 2009. Outside these years, node A was the prominent node according to kiin. Out
degree (kiout) also identified the key player to be node A, with the exception of 2007 where an offshore
island, Tiritiri Matangi (node J) was a source of output into the network.
Year In degree (kiin)a Out degree (ki
out)b Betweenness (Bi)c 2007 Burwood Bush breeding centre Tiritiri Matangi Island Burwood Bush breeding centre,
Mana Island, Maud Island 2007 to 2008 Murchison Mountains Burwood Bush breeding centre Burwood Bush breeding centre 2007 to 2009 Murchison Mountains Burwood Bush breeding centre Burwood Bush breeding centre 2007 to 2010 Burwood Bush breeding centre Burwood Bush breeding centre Burwood Bush breeding centre 2007 to 2011 Burwood Bush breeding centre Burwood Bush breeding centre Burwood Bush breeding centre Table 2-2 Identification of key takahe population location within the annual translocation networks, categorised by the highest node-based measure.
When plotting kiin against ki
out, apart from 2007 when measures were less defined between nodes in
the network, the cumulative networks identified node A every year as the hub of the network (Figure
2-2). Likewise, the wild Murchison Mountains population (node Q) is consistently the sink of the
network with regularly high kiin and low ki
out scores (Figure 2-2). Although Mana Island (node C) has
been in the network since 2007, its kiout steadily increased, whereas its ki
in remained constant. The
emergence of node C as a potential source within the network becomes apparent from 2010 onwards,
and in the final 2011 network its kiout is 3 times that of its ki
in score. An additional note is the
increasing importance of Wildbase Hospital (node G). Although not the most obvious hub of the
network, its kiin and ki
out scores increased proportionately and steadily.
Identifying key players according to the highest measure of weighted betweenness (Bi) produced a
consistent key player (Figure 2-1, Table 2-2). Node A had the highest measures every year from 2007
to 2011 when previous connections are included. In 2007, Bi measures were more conserved among
nodes in the network. Therefore, three nodes, A, C, and D (Maud Island) were identified as key
players.
38
Figure 2-2 Relationship between weighted in degree (kiin) (animals moving into a location) and weighted out degree (ki
out) (animals moving out of a location) for nodes (i.e., isolated geographic locations used for takahe conservation) of the cumulative (a) 2007, (b) 2007–2008, (c) 2007–2009, (d) 2007–2010, and (e) 2007–2011 final network of takahe translocations. Alphabetical labels are as for Figure 2-1
39
2.5. Discussion
Conservation of takahe populations has been of significant interest internationally. However, disease
has rarely been questioned as a contributing factor in the decline of the population, with only one
retrospective investigation highlighting the lack of cohesive disease reports and post-mortem
protocols (McLelland et al. 2011). As a consequence of past events, the population has declined to a
small number reliant on active management and translocations (Hegg et al. 2012). The translocation
database of this flightless endangered species provides an accurate record of locations, individuals,
and movements because there is no natural immigration or emigration between the small remnant
populations. The ability to model movements of takahe for the last five years has provided the
opportunity to construct complete networks and draw valid conclusions for the timespan available.
This is often not the case when analysing movements of animals. Data are often missing or unknown,
and ascertaining all associations within a country is rarely possible (Ortiz-Pelaez et al. 2006).
Therefore, range-restricted endangered animals, such as takahe, are ideal candidates for SNA. This
analysis can then be used to determine where to target disease surveillance or to control the spread of
disease in an outbreak situation.
2.5.1. Application of network analysis to takahe movements
Approaching the analysis of the takahe database using multiple network measures allowed
comparison of network methods and subsequent outcomes. Annual networks were constructed to
include translocations from previous years on record because takahe are a long-lived flightless
species, and when an individual is moved to a new location, it cannot emigrate to another area without
human intervention. Therefore, although the translocation may have occurred several years before the
year of analysis, reconnection of the location to the network poses the risk of transmission of
pathogens from the extant population back into the network. Annual analysis of the translocation
network informs management and identifies shifts in key nodes of contact for the species over time
(Figure 2-2, Table 2-2). The final 2007–2011 cumulative network was most important in the
assessment of the current state of the network in terms of identifying suitable locations for disease
surveillance (Figure 2-1, Figure 2-2).
40
The number of locations and interactions between these sites increased annually over the time frame
of the study (Table 2-1), correlating with the increasing use of translocations as an endangered species
management tool worldwide (Fischer & Lindenmayer 2000) and in New Zealand (Sherley et al.
2010). As more locations are added to the network, the likelihood of introduction and exposure to
infectious agents may increase due to increased connectivity, and potential clustering between
populations. Infectious diseases spread more easily in small-world networks than in regular and
random networks. Disease spread is unpredictable in small-world networks due to the high degree of
local clustering encountered (Jeger et al. 2007). Limitations of the takahe data, in terms of sample size
and lack of complexity in the contact networks, restrict the ability to confirm the presence of small-
world attributes in the takahe networks. However, there appears to be a trend for the annual networks
to fluctuate and display some attributes of small-world characteristics, with the most obvious being in
the final 2007–2010 network where the Lav/CCav was substantially smaller than that of the random
equivalent networks (Table 2-1). From this it could be inferred that as translocations continue over
time and with the establishment of new connections, small-world characteristics may develop. The
risks associated with small-world characteristics in terms of ease of pathogen spread between
locations should be kept in mind when considering the translocation of birds between populations.
2.5.2. Identification of hubs, sinks, and sources
Detecting wildlife disease trends requires optimal diagnostic sensitivity and specificity, appropriate
sample sizes and sampling strategies, and analytical tools that can identify anomalies in space and
time in order to inform decision making. However, sampling by its nature is only representative of a
small proportion of a real system. Surveillance programs require appropriate sampling to provide a
good representation of a population network. Interpreting connectivity within a population can
provide insights into the dynamics of a system, identifying players that are most influential. This
study used the takahe translocation database as a model to predict which locations may act as hubs,
sinks, and sources of pathogens.
A notable result of this investigation is the central position of a captive breeding centre (A), wildlife
hospital (G), and wild population (Q) within the translocation network. When looking at weighted
centrality measures of degree (kiin and ki
out) and betweenness (Bi), although these measures
41
occasionally identified different key players within a network, consistently locations A, G, and Q
demonstrated important roles in connectivity between locations (Table 2-2).
Rediscovery of a few wild takahe in the Murchison Mountains in 1984 led to the translocation of birds
to and from the wild to sanctuaries across New Zealand (Jamieson & Ryan 1999). The wild
population soon became a source for genetic input into the captive populations; however, a large
proportion of takahe are still resident in the Murchison Mountains (Hegg et al. 2012). Our networks
illustrate the importance of island sanctuaries, with island locations being identified as key players
according to kiout and Bi measures in 2007. After 2007, management increased the role of the wild
population in the network. In the timespan included in this investigation, the Murchison Mountain site
received many takahe from external sources with relatively little emigration of individuals (Figure
2-2). Management decisions made in 2009 and 2010 to treat the wild birds in the Murchison
Mountains as a separate population led to the cessation of relocations to and from Fiordland (Glen
Greaves, personal communication). In the network analysis, this is represented by stagnation of
centrality measures for this node from 2009 onwards (Figure 2-2). A location such as the Murchison
Mountains, which has had multiple immigration events with little emigration, could act as a sink for
pathogen evolution within that population; the birds could carry unique microflora that could be
pathogenic to naïve-introduced animals.
Captive breeding facilities, such as Burwood Bush breeding centre (A), often have a high turnover of
animals and intensive production of stock for reintroduction to the wider population. A common
difficulty with endangered species, such as takahe, is that poor breeding success leads to the need for
mixing of mates to find compatible couplings (Jamieson et al. 2006). Additionally input of new
breeding lines from external sources is required to minimize inbreeding (Jamieson & Ryan 2001). As
such it is not surprising that Burwood Bush breeding centre has emerged as a substantial hub of the
takahe network, irrespective of management decisions with regards to other reserves. Consistently, the
location increased its connections to and from multiple locations (Figure 2-2) and played a major role
as an intermediary between sites, as shown by the Bi and ki measures (Table 2-2). With such
connected locations containing animals originating from multiple locations, focusing effort and
surveillance within a hub provides insight into what is flowing through the entire network.
42
Wildlife health treatment centres, such as Wildbase Hospital (G), are of a different nature than the
mainland and island reserves included in the network. The number of takahe sent to veterinary
hospitals is likely to be variable depending on the health status of the population, whereas movements
between other locations are controlled by management decisions. The veterinary hospital (G)
appeared to gain importance over time where the number of birds transiting through the location
increased, as shown by growth in kiin and ki
out scores (Figure 2-2). The hospital was also an
intermediary location, as shown by high measures of Bi, particularly evident in the final 2011 network
(Figure 2-1). Animals sent to a hospital are more likely to be ill or injured. Therefore, the animals may
be more vulnerable to pathogen invasion and could present a significant disease risk by acting as
carriers or vectors of new diseases to wild populations when released (Viggers et al. 1993).
A translocation database is by no means static over time; locations may be included one year and not
in subsequent years. Establishment of new sites in the network for advocacy or display purposes, such
as Peacock Springs wildlife park (L) and Willowbank reserve (O), have had inputs of birds into them
but are not connected to other locations (Figure 2-2). Sites such as these are regarded as peripheral
and relatively isolated and could be targeted to determine extent of disease spread. If reconnected to
the network by the process of translocation of takahe from these nodes to a hub, such as Burwood
Bush breeding centre, this could create a transmission pathway to introduce a novel strain to the hub
that could subsequently spread throughout the network. If the networks were to be analysed annually
without accounting for historic translocations, peripheral sites would drop out from the networks. If a
translocation were to occur either to or from this site in the future, we would not know the risk that the
birds already in the location may pose due to previous connections from historic translocations.
2.5.3. Limitations
Like most models of complex biological systems, the SNA produced here has limitations and further
refinements are required to reflect real situations. Records used in this analysis span only five years of
a 30 plus year history of population management due to the scarcity of recorded data prior to 2007.
Limitations in the collation of data for this investigation demonstrated the need for conservation
managers to maintain accurate records of individual takahe movements and to file this information in
one central database. Seasonal timing of translocations and the numbers of birds translocated at a
43
given time are components that may influence the importance of a location in the network due to
differing influences on the likelihood of transmission and susceptibility to pathogens. Relationships
among pathogens, seasonal host infection, and transmission are currently poorly understood. This
information would add another level of complexity and increase validity of network associations.
Island sanctuaries, both mainland and offshore, used for takahe conservation are also used as predator
free reserves for other threatened New Zealand native fauna and flora (Sherley et al. 2010). The
introduction of any animal to a new site may expose an immunologically naive animal to potentially
pathogenic organisms residing in other species (Cunningham 1996). Many of the sites have histories
of translocations of other species, which could interact with takahe and transmit infective agents to
cohabiting takahe. Integration of translocation databases of all introduced species within the takahe
reserves would provide a better understanding of the reserve networks and connectivity beyond that of
just takahe. Disease surveillance could then be assigned according to complete ecological networks.
However, if cross-species transmission from resident and migratory species already present in
reserves is considered, then the complexity of the network could rapidly increase beyond our ability to
describe it. In addition to takahe, people and field equipment associated with a translocation are
moving between locations within the network. This could increase the connectivity between
previously isolated locations. To compensate for this possibility, it is recommended that strict
quarantine and biosecurity checks continue to be practiced prior to and on arrival at these locations.
2.5.4. Conservation implications and future directions
SNA of historical takahe translocations enabled us to characterize key components of the conservation
management network that are potentially important in pathogen transmission. Our findings provide
advice for takahe conservation management, where networks are used as component of decisions. It
could be applied to help reduce the likelihood of disease spread, identify target locations to monitor
for disease emergence and inform mitigation actions should an outbreak occur.
Disease screening can be a very costly component of a translocation, and epidemiologists have been
limited in gaining sufficient sample sizes to provide informative meaningful results (Gartrell et al.
2006). Additionally, choosing influential pathogens or those which are relevant to the target species
44
can be a difficult task. Mathews et al. (2006) suggest that initial screening should be as broad as
possible, using existing knowledge of related species. Pooled sampling is often proposed as a low-cost
solution (Mathews et al. 2006). However, it is less informative than individual testing. This study
identified key locations within the takahe network to impose disease surveillance if funds were
limited and management was unable to monitor an entire population. Locations can be selected on the
basis of potential threats and / or being representative of the takahe population as a whole.
Optimisation of sample collection, sampling a broad range of pathogens from individuals in key
locations identified from network analysis, would allow economically viable targeted monitoring of
potential threats to a population and inform normal ranges of pathogen prevalence, thus providing
epidemiologically sound advice on disease management.
Conservation managers could use a translocation network, like the one used in this study, to inform
decisions when moving animals between locations. Takahe movement network analysis has
highlighted that historic decisions of introducing intensively reared birds from a breeding centre into
the one remaining wild population in the Murchison Mountains may have exposed a wild population
to exotic pathogens, which in the face of disease threats may be a risky strategy. Therefore, before
commencing a translocation from one location to another, it would be worthwhile noting the current
network measures of the existing and new location. If the translocation would reconnect a peripheral
location, a sink, to a hub within the network, informed decisions on quarantine procedures and
additional disease screening could be made.
A retrospective network analysis of the 2001 foot and mouth disease outbreak in the United Kingdom
(Ortiz-Pelaez et al. 2006) uncovered how network analysis could inform mitigation to prevent spread
of epidemics. If an infectious disease outbreak were to occur within the takahe population, sites with
the highest centrality measures within a network could be isolated and movement restrictions imposed
to most effectively disrupt transmission pathways.
Targeted and opportunistic disease surveillance, focusing on common pathogens and their respective
strain types, provides detailed information on pathogen diversity and relative importance of aetiology
of disease in the host (Smith et al. 2009). Comprehension of the diversity and abundance of pathogens
in a community and ecosystem will increase our grasp on potential transmission and evolution of
45
pathogens within the species of interest (Thompson et al. 2010). In addition, it would allow one to
trace and overlay a pathogen network with a translocation network and gain insight into the
epidemiology of pathogens within an altered ecosystem.
2.6. Acknowledgments
This study was been funded by the Allan Wilson Centre for Molecular Ecology and Evolution. I
would like to thank Wildbase Hospital, P. Marsh, G. Greaves, and L. Kilduff from the Department of
Conservation for the provision of data, and N. Nelson, K. McInnes, L. Howe, and K. S. Richardson
for comments on the manuscript. I acknowledge the time and insight provided by anonymous
reviewers whose comments improved this manuscript.
2.7. Supporting Information
Map of takahe locations within New Zealand (Appendix 9.2-1), takahe movement data (Appendix
9.2-2), and individual node measures (Appendix 9.2-3) are available in the appendix.
46
47
48
CHAPTER 3
USING A COMMON COMMENSAL BACTERIUM IN
ENDANGERED TAKAHE (PORHYRIO HOCHSTETTERI), AS A
MODEL TO EXPLORE PATHOGEN DYNAMICS IN ISOLATED
WILDLIFE POPULATIONS
49
50
3. Using a common commensal bacterium in endangered takahe
(Porphyrio hochstetteri), as a model to explore pathogen dynamics in
isolated wildlife populations
3.1. Abstract
Predicting and preventing outbreaks of infectious disease in endangered wildlife is problematic
without an understanding of the biotic and abiotic factors that influence pathogen transmission, and
the genetic variation of microorganisms within and between these highly modified host communities.
Using a common commensal bacterium, such as Campylobacter spp., in endangered takahe
(Porphyrio hochstetteri) populations enables the development of a model to study pathogen dynamics
within isolated wildlife populations which are connected through ongoing translocations. Takahe are
an endangered flightless bird endemic to New Zealand, with a population of approximately 230
individuals. Insurance populations were founded from a single remnant wild population and have
been established within multiple reserves. Management has resulted in the formation of several
fragmented sub-populations maintained and connected through regular translocations. This
investigation tested 118 takahe from eight locations for faecal Campylobacter spp. via culture and
DNA extraction, with species assignment conducted by PCR. Factors relating to population
connectivity and host life history were explored using multivariate analytical methods to determine
associations between host variables and bacterial prevalence. The apparent prevalence of
Campylobacter spp. carriage in takahe was 99%, one of the highest reported in avian populations.
Variation in prevalence was evident between Campylobacter spp. identified; C. sp. nova 1 (90%)
colonised the majority of takahe tested. Prevalence of C. jejuni (38%) and C. coli (24%) was different
between takahe sub-populations, and this may be explained by factors related to population
management, captivity, rearing environment and the presence of agricultural practices in the location
in which birds were sampled. Modelling of a commensal microorganism within takahe meta-
populations suggests that anthropogenic management of endangered species within altered
environments may have unforeseen effects on microbial exposure, carriage and disease risk.
Translocation of wildlife between locations could have unpredictable consequences including the
spread of novel microbes between isolated populations.
51
3.2. Introduction
The inadvertent introduction or emergence of infectious disease through translocation of animals into
new populations or ecosystems is a major concern when considering management of wildlife
populations (Thompson et al. 2010). Translocation of individuals between previously isolated
ecosystems removes barriers to the exchange of pathogens (Power et al. 2013). These translocations
can potentially transfer exotic pathogens into extant populations with no effective immunity at the
release site (Anderson & May 1986; Woodford & Rossiter 1993). Threats posed by management
actions are particularly pertinent to endangered species that are maintained in fragmented isolated
populations and are heavily reliant on conservation measures for population viability.
Epidemiological investigations attempt to understand the roles biotic and abiotic factors have upon the
transmission of pathogens to inform risk assessments and aid in the prevention and control of
outbreak of disease. This study used a common commensal bacterium Campylobacter spp. as a model
to explore host-microbial dynamics in populations of the endangered flightless takahe (Porphyrio
hochstetteri) in New Zealand.
Human mediated movement of endangered animals through the process of translocation has
developed into a commonly-used conservation tool worldwide (Fischer & Lindenmayer 2000; Griffith
et al. 1989). In New Zealand, over 900 known translocation events of native terrestrial animals have
occurred in the last 70 years (Sherley et al. 2010). The takahe is an endemic New Zealand flightless
rail that was thought to be extinct until a small population was rediscovered in Fiordland, New
Zealand in 1948 (Ballance 2001). To date a wild Fiordland population remains, with additional
insurance populations located in predator-free offshore and mainland reserves (Wickes et al. 2009).
Insurance populations are descendants of the original Fiordland population, and as such are actively
managed to minimise the potential detrimental effects of inbreeding (Jamieson et al. 2006) by means
of captive breeding and multiple translocations per annum (Grange et al. 2014). Sub-populations are
geographically isolated and natural dispersal of takahe is not possible. Microbial carriage is predicted
to differ between host populations maintained within a meta-population due to variation in social
connectivity, the environment and life history (VanderWaal et al. 2013b).
Social network analysis of a takahe translocation database identified a complex network of sub-
populations which were predicted to vary in their likelihood of maintaining and transmitting
52
pathogens (Grange et al. 2014). Highly connected takahe populations may have an important role in
pathogen dispersal, whereas groups with fewer translocations could act as sinks or sources of exotic
pathogens due to an increased possibility of allopatric speciation after a period of isolation (Grange et
al. 2014). The work by Grange et al. (2014) provided a basis for an empirical investigation into the
molecular epidemiology of infectious organisms in the fragmented takahe population. The aim of this
study was to investigate the effects of population isolation, management, host biotic factors and
environmental variation on the carriage of infectious organisms. The outcomes of epidemiological
models could inform disease risk assessments for the management of fragmented endangered wildlife
populations.
3.3. Materials and methods
3.3.1. Study population
Approximately half (n = 118) of the total population of takahe were opportunistically tested for faecal
carriage of Campylobacter spp. from eight locations (Table 3-1) within New Zealand between: March
and April 2012 (sampling period 1, n = 71), August and November 2012 (sampling period 2, n = 8),
and February and April 2013 (sampling period 3, n = 39). Data from individual takahe, identified by
Department of Conservation certified band numbers included: age, sex, nest site location, rearing
method (wild, puppet (Eason & Willans 2001) or fostered) and location at time of sampling. The two
variables age (9 unknown) and sex (22 unknown) were incomplete due to restricted monitoring of the
wild population.
3.3.2. Sample collection
Samples were collected opportunistically during pre-translocation disease screening or annual health
checks of takahe. Faecal samples were collected during handling and placed into sterile sealed
containers. Sterile swabs were immediately inserted into the fresh faecal sample and stored in Aimes
charcoal transport media (Copan, California, USA). All samples were transported refrigerated and
stored at 4oC for 1 to 7 days prior to culture and faecal DNA extraction.
53
3.3.3. Microbiological culture and DNA extraction
Swabs were suspended in 2ml Phosphate Buffered Saline (PBS, pH 7.3). 100µl of inoculated PBS
was enriched in 2ml Bolton's broth (Lab M, Bury, England) for 48 hours at 42oC in microaerobic
conditions. Bolton’s broth was sub-cultured onto modified Charcoal Cefoperazone Deoxycholate
Agar (mCCDA) (Fort Richard, Auckland, New Zealand) and incubated for 48 hours at 42oC in
microaerobic conditions. Two colonies from each Campylobacter spp. positive mCCDA plate were
chosen at random and sub-cultured onto Columbia horse blood agar plates (Fort Richard, Auckland,
New Zealand). Plates were incubated for 24 hours at 42oC in microaerobic conditions. DNA
extraction from pure cultures was conducted after 24 hours using a 2% Chelex 100 resin (BioRad,
Auckland, New Zealand) suspension in sterile MilliQ water (Merck, Palmerston North, New Zealand)
and boiling at 100oC for 10 minutes, followed by extraction and storage of the supernatant at -20 oC.
Faecal samples were subject to direct DNA extraction using modified stool pathogen detection
protocols, outlined in the QiAmp Stool Minikit (Bio-Strategy, Auckland, New Zealand). Initial
incubation was at 95oC and elution was in 100µl sterile milliQ water (Merck, Palmerston North, New
Zealand). DNA was stored at -20 oC until required.
3.3.4. Molecular confirmation and speciation
Suspected Campylobacter spp. isolates and faecal DNA were subject to polymerase chain reaction
(PCR) using previously described primers and protocols for identification of the mapA gene in
Campylobacter jejuni (Mullner et al. 2010; Stucki et al. 1995) and the ceuE gene in Campylobacter
coli (Denis et al. 2001; Gonzalez et al. 1997).
Samples were tested for a newly identified putative species of Campylobacter, named Campylobacter
species nova 1 (French et al. 2014). Despite extensive examination of multiple hosts, including
farmed livestock and wildlife, C. sp. nova 1 has only been identified in New Zealand surface water
and members of the Rallidae family (French et al. 2014). Identification of C. sp. nova 1 was
conducted using an in house PCR that targets a short section of a putative C4-dicarboxylate trans-
membrane transport gene believed to be found only in C. sp. nova 1. Forward Aot10724 5’
GGTGTGTTTGCTGGTCTTGTATTGGC 3 and reverse Aot10724 5 AAATCCACTCCCCGTTT
TGCGA 3’ primers were designed using Geneious software v6.1 (Drummond et al. 2013), and
compared to the GenBank database, where no matches were found. The expected product size was
54
106 base pairs. PCR conditions for C. sp. nova 1 were as follows: 95oC for 2 minutes for initial
denaturation, followed by 40 cycles of 94oC for 20 sec, 55oC for 20 sec and 72oC for 10 sec. The 20
μL PCR reaction mix consisted of 2μL 10x PCR buffer, 0.4µL magnesium chloride (1 mM), 1µL
dNTPs (200 μM per dNTP), 2μL of each primer (4pmol), 1 unit Platinum Taq DNA polymerase (Life
Technologies, Auckland, New Zealand) and 2μL DNA at 40ng.
Isolates and faecal DNA samples which tested negative for the three Campylobacter spp. described
above were subject to a PCR targeting a region of the 16S rRNA gene for genus level confirmation of
unknown Campylobacter spp. (Linton et al. 1997). Each PCR reaction was run with positive control
identified by whole genome sequencing and a negative control of water. To confirm successful
amplification, all PCR products were run on a 1% gel agarose in Tris-borate-EDTA (TBE) buffer
followed by staining with ethidium bromide and visualised by exposure to UV light.
3.3.5. Prevalence of Campylobacter spp. in takahe
3.3.5.1. Apparent prevalence
Detection of Campylobacter spp. in takahe was dependent on the accuracy of test diagnostics. There
was no gold standard method for culture of Campylobacter spp. This investigation used two methods
of testing, culture and faecal DNA extraction both followed by PCR. By combining the results of the
two tests, an individual was classified as positive if it was positive for either test. This method reduced
the likelihood of false negatives which may be encountered during culture and selection of
Campylobacter spp. isolates. Apparent overall prevalence estimates of Campylobacter spp. were
calculated from the combined test outcomes in R software (R Core Team 2013) using the EpiR
package (Stevenson 2014). A Venn diagram was created in R software (R Core Team 2013) using the
VennDiagram package (Chen 2013) to show single and multiple carriage of Campylobacter spp.
3.3.5.2. Estimates of true prevalence, sensitivity and specificity of tests
Markov Chain Monte Carlo (MCMC) latent class analysis (LCA) was conducted (Appendix 9.3-1) in
R software (R Core Team 2013) to generate estimates of the true prevalence, and the sensitivity and
specificity of the two tests in three populations, assuming that the tests were conditionally
independent. The LCA methods were based on those created for analysis of imperfect diagnostic tests
in the absence of a gold standard (Branscum et al. 2005). The routine allows posterior testing for
differences in prevalence estimations between populations. The apparent prevalence estimates of the
55
three Campylobacter spp.; C. jejuni, C. coli and C. sp. nova 1 for the three management populations
(wild, breeding and insurance) (Table 3-1) were used for input into the LCA model. Prior estimates of
test sensitivity and specificity and estimates of prevalence were derived from expert opinion and were
specified using beta distributions calculated in Betabuster software (Chun-Lung 2014). Stability of
each model was tested using uninformative priors (Beta(1,1)) and posterior distributions were
compared to models with informative priors. Pairwise tests for differences in prevalence of C. coli, C.
jejuni and C. sp. nova 1 between the three host populations were performed using a Bayesian
statistical probability for estimating differences in prevalence (Pr) between populations, where values
close to 0 and 1 indicate potential significant differences. Plots of true and apparent prevalence
estimates with 95% confidence intervals were created in the ggplot2 package (Wickham 2009).
3.3.6. Exploratory analysis of explanatory covariates
3.3.6.1. Allocation of categorical explanatory covariates
All explanatory variables were categorical and assigned as described below and in Table 3-1. Age of
the takahe was categorised as juvenile below the age of two, and adult above the age of two. Takahe
sex was categorised as male, female or unknown if genetic sex identification had not been performed.
Temporal trends were assessed by the comparison of three sampling periods described previously.
Location and nest site variables were analysed in multiple formats according to physical location (NZ
island, geography, proximity to farming) as well as assignments based on type of location
management and connectivity (in degree, out degree, betweenness) (Table 3-1). New Zealand was
divided into two categories based on the two major land masses of New Zealand (North Island/South).
Geography (mainland/island) was classified according to the location being on the mainland of New
Zealand or an offshore island. Conservation management of takahe differs between locations, with
three management types being evident (breeding/wild/insurance). The wild population are free
ranging in their original location, insurance populations are the translocated populations located in
reserves, and the breeding centre acts as a source of birds for the insurance populations. Proximity to
farming (farmed/remote) was assigned according to whether the location was directly surrounded by
agriculture. An additional variable, translocation status, was measured on an individual level based on
whether the animal’s location at time of sampling differed from its nest site (location the takahe egg
was laid).
56
Location * No. of takahe NZ island Geography Management Proximity to farming In Degree# Out Degree# Betweenness# Burwood Bush 34 South Mainland Breeding Farmed High High High Kapiti Island 0 North Island Insurance Remote NA NA NA Mana Island 6 North Island Insurance Remote Low High Medium Maud Island 2 South Island Insurance Remote Low Low Low Maungatautari reserve 6 North Mainland Insurance Farmed Low Low Low Murchison Mountains 44 South Mainland Wild Remote High Low Low Private Island 10 South Island Insurance Remote Low Low Medium Te Anau Reserve 2 South Mainland Insurance Farmed Low Low Medium Tiritiri Matangi Island 13 North Island Insurance Remote Low Low Low Willowbank Reserve 1 South Mainland Insurance Farmed Low Low Low *Location refers to either location at time of sampling or nest site (where egg was laid) #Variable only applicable to locations where sampling of takahe occurred, terms and values are in
Chapter 2
Table 3-1 Locations, number of takahe included in the analysis and corresponding categorical variables for input into multiple correspondence analysis, latent class analysis and logistic regression modelling
57
The effect of population connectivity on prevalence of Campylobacter spp. was explored using social
network measures. Location weighted betweenness, in degree and out degree measures were derived
from social network analyses previously conducted on a historic (2007 to 2011) takahe translocation
database (Grange et al. 2014). Weighted betweenness (Opsahl et al. 2010) is a measure of centrality
that represents the likelihood of a takahe population connecting two random populations via the
shortest path, accounting for the sum of translocations in and out of the population. Weighted degree
(Opsahl et al. 2010) was calculated from the number of direct connections between a location and its
neighbour, and the sum of translocations between those connections. Degree measures can be divided
into in and out measures which represent the connections into and out of a population respectively. In
and out degree measures were categorised as low below a measure of 10 and high above 10.
Betweenness was broken into three categories: low (0), medium (9-19) and high (78).
3.3.7. Multiple correspondence analysis
Multiple correspondence analysis was used to explore interactions between the categorical variables
and the apparent prevalence of the three identified Campylobacter spp. MCA detects and represents
underlying structures in a data set by presenting data as points in a low-dimensional Euclidean space
(Greenacre 2007) and was used as an exploratory tool prior to statistical modelling. MCA and
hierarchical analysis was performed in R software (R Core Team 2013) using the factoMineR and
ggplot2 packages (Husson et al. 2013; Wickham 2009). Values from the first two dimensions, which
explain the greatest amount of the variation in the data (Greenacre 2007) were plotted. Graphical
overlapping and close proximity between data points indicates a close relationship between variables
(words) and individuals (data points). Hierarchical analysis was performed on the outputs of the MCA
to determine which variables contributed the most to the clustering of data points.
3.3.8. Multivariate logistic regression modelling
Explanatory variables were analysed using univariate and multivariate logistic regression in R
software (R Core Team 2013). Generalised linear models with binomial errors were used to identify
variables associated with the prevalence of C. jejuni, C. coli and C. sp. nova 1, unidentified
Campylobacter spp. and multiple Campylobacter spp. carriage in takahe. All biologically plausible
variables were included in the analysis. However, with limited data available, univariate screening
was employed to reduce the number of variables included in the multivariate analyses to those which
58
had a more direct impact on the outcome. Explanatory variables in univariate analysis with ANOVA
test Chi-squared p-values of <0.20 when compared to the null model were included in an initial
multivariate model (Appendix S6-8). Two-way interactions between univariate explanatory variables
were explored, however none met the criteria (p-value <0.05) to be included in the final model.
A process of model simplification was used to create a parsimonious final model with variables of
significance associated with the carriage of Campylobacter spp. Stepwise elimination of variables of
least significance was conducted based on Akaike’s Information Criterion (AIC) scores. Variables
were retained in the model if they showed confounding effects and improved model fit, even if they
were not significant. Odds ratios and corresponding 95% confidence intervals were calculated for the
categorical variables included in the final model. Significance testing between variable levels was
conducted using Wald’s tests. The goodness of fit was assessed using the Likelihood ratio and
Hosmer-Lemeshow tests.
Two multivariate models were created for each Campylobacter spp. Each model contained variables
identified as important from univariate analyses. One model was based on a reduced dataset
containing age or sex, if they were significant in univariate ANOVA analyses. The second model used
the full dataset, excluding age or sex. If the age or sex were not significant in the final model, the
multivariate model using the full dataset was accepted. If age or sex were significant in the final
multivariate model, the two final models were compared and the best fitting model accepted using
AIC.
3.4. Results
3.4.1. Apparent prevalence of Campylobacter spp. in takahe
A total of 117 takahe tested positive for Campylobacter spp. via either culture or DNA testing of
faecal material, giving an overall (based on the results of the two tests) apparent prevalence of 99.2%
[95% CI, 95.4 to 100%]. Three Campylobacter spp. were identified: C. jejuni (apparent prevalence,
AP, 38.1% [95% CI, 29.4 to 47.5%]), C. coli (AP 23.7% [95% CI, 16.4 to 32.4%]) and C. sp. nova 1
(AP 89.8% [95% CI, 82.9 to 94.6%]) (Figure 3-1). Campylobacter spp. belonging to the
Campylobacter genus were also detected and remain unidentified (AP 5.9% [95% CI, 2.4 to 11.8%]).
Multiple carriage of Campylobacter spp. was detected, with just over half (50.9%) of takahe testing
59
positive for two or more Campylobacter spp. Most commonly combinations of C. sp. nova 1 / C.
jejuni (29/118) and C. sp. nova 1 / C. coli (18/105) were detected in takahe faeces (Figure 3-2).
Figure 3-1 Apparent and true prevalence with 95% confidence intervals of Campylobacter sp. nova 1, Campylobacter jejuni and Campylobacter coli in populations of takahe (Porphyrio hochstetteri)
3.4.2. Estimates of true prevalence using imperfect tests
True prevalence estimates for the three identified Campylobacter spp., C. jejuni, C. coli and C. sp.
nova 1, within the three takahe management subpopulations are presented in Figure 3-1. LCA model
outputs including estimates of true prevalence, probability of differences in prevalence between
population pairs (P), and the true sensitivity / specificity of the tests are detailed in Appendix 9.3-2.
There was no significant difference in the prevalence of C. sp. nova 1 in the pairwise comparisons of
the breeding, wild and insurance populations. The prevalence of C. coli and C. jejuni in the wild
population was significantly different from the breeding (C. coli P = 0.001, C. jejuni P = 0.02) and
insurance (C. coli P = 0.009, C. jejuni P = 0.035) populations, with no significant difference observed
between the breeding and insurance populations (C. coli P = 0.292, C. jejuni P = 0.334). Sensitivity
and specificity for the two tests did not differ largely across the three species of Campylobacter;
however there was some variation between the types of tests used. DNA extraction followed by PCR
60
was more accurate than culture for diagnostic testing, with sensitivity/specificity estimates for C.
jejuni of 91.4/88.3%, C. coli 92.5/90.7% and C. sp. nova 1 97.8/88.6%. Culture followed by PCR was
highly specific but less sensitive than DNA extraction with sensitivity/specificity estimates of C.
jejuni 62.9/95.7%, C. coli 65.3/97.7% and C. sp. nova 1 69.1/97.8%.
Figure 3-2 Venn diagram of Campylobacter spp. carriage in 117/118 takahe (Porphyrio hochstetteri). Numbers represent individuals positive for each Campylobacter species, overlapping areas represent individuals carrying combinations of Campylobacter spp.
3.4.3. Analysis of explanatory covariates associated with the carriage of Campylobacter spp.
Takahe showed some variability in covariates associated with carriage, depending on the species of
Campylobacter concerned. When investigating individual takahe covariate patterns, three groups or
clusters are evident, indicating similarities of takahe profiles within but not between the clusters
(Appendix 9.3-3A). According to hierarchical analysis, population management (Appendix 9.3-3B)
and betweenness (Appendix 9.3-3C) contributed most to the separation of the observations into
groups. In addition, whilst carriage of C. sp. nova 1 does not fall within any of these groups due to the
very high prevalence, carriage of C. jejuni and absence of C. coli both fall within the wild population /
low betweenness cluster, indicating some correlation between these variables (Appendix 9.3-3A).
Aggregation of variables within the MCA plot indicate some co-linearity may be present (Appendix
9.3-3) and could be influential on multivariate analysis outcomes, but confounding variables were
retained if they improved model fit.
61
The high prevalence of C. sp. nova 1 and low prevalence of the unidentified Campylobacter spp. in
takahe tested limits the analysis of variables which may be associated with the presence or absence of
those organisms. Univariate and multivariate analysis found no significant association between the
explanatory variables and carriage of C. sp. nova 1 (Appendix 9.3-5, Table 3-2) or unidentified
Campylobacter spp. (Appendix 9.3-7, Table 3-2).
The final multivariate model for C. coli was based on the reduced dataset (96 individuals) due to the
significant effect of age on the carriage of C. coli in multivariate analyses. Juveniles were more likely
(p-value = 0.026) to test positive for C. coli than adults. Takahe born in the breeding centre were
significantly less likely (p-value = 0.04) to have C. coli than those born in an insurance population
(Table 3-2), but there was no significant difference between the breeding and wild populations (p-
value = 0.3). Being located within close proximity to farming was significantly associated with
carriage of C. coli (p-value = 0.001) (Table 3-2).
Takahe located in the wild founder population were significantly more likely to carry C. jejuni (p-
value = 0.02) than the insurance and breeding populations (p-value = 0.003) (Table 3-2). Although no
significant differences were observed in the carriage of C. coli between the levels of nest site
management (Table 3-2), inclusion of the variable improved model fit. Carriage of C. jejuni was
associated with sampling period, with a significantly lower prevalence observed in the third sampling
period (Feb-Apr 2013), compared with the first (Mar-Apr 2012) (p-value = 0.003).Multiple carriage
(more than 1 species) of Campylobacter spp. was apparent in a subsection of the population with
some strong associations with nest site. Takahe born in locations within close proximity to farming
were significantly more likely to carry multiple species than those born in remote locations (p-value =
0.02) (Table 3-2).
62
Test species Variable Level Coefficient (SE) Odds ratio (95% CI) p-value *C. colia Intercept Insurance / Remote from farms / Adult - - - Nest site Breeding -1.57 (0.77) 0.21 (0.04-0.86) 0.041 Wild -0.69 (0.72) 0.50 (0.11-1.97) 0.338 Location Close to farms 2.42 (0.65) 11.22 (3.41-45.55) <0.001 Age Juvenile 1.39 (0.63) 4.03 (1.23-14.83) 0.026 C. jejunib Intercept Period 1 (Mar-Apr 2012) / Insurance / Insurance - - - Sampling period Period 2 (Aug-Nov 2012) 0.23 (0.86) 0.43 (0.15-1.11) 0.787 Period 3 (Feb-Apr 2013) -1.59 (0.53) 1.26 (0.21-6.77) 0.003 Nest site Breeding 1.07 (0.75) 0.20 (0.06-0.55) 0.153 Wild -0.32 (0.77) 2.92 (0.69-13.56) 0.683 Location Breeding -0.35 (0.76) 0.73 (0.15-3.32) 0.642 Wild 1.95 (0.81) 7.02 (1.49-37.33) 0.016 C. sp. nova 1c Intercept Remote from farms - - - Nest site Close to farms -1.42 (1.07) 0.24 (0.01-1.33) 0.183 Unidentified Campylobacter spp.d Intercept No / High - - - Translocated Yes 1.17 (0.80) 3.22 (0.66-17.41) 0.145 In degree Low 1.09 (0.80) 2.97 (0.61-16.09) 0.175 Multiple carriagec Intercept Remote from farms - - - Nest site Close to farms 0.45 (1.08) 2.93 (1.24-7.41) 0.018 *model based on a smaller dataset due to 22 missing data for age Likelihood ratio test: a p <0.001 d.f = 4, b p <0.001 d.f = 6, c p=0.113 d.f =1, d p=0.148 d.f = 2, e p = 0.01 d.f = 1
Table 3-2 Multivariate generalised linear model showing variables of significance for the carriage of Campylobacter jejuni, Campylobacter coli and Campylobacter species nova 1 in takahe (Porphyrio hochstetteri) faeces.
63
3.5. Discussion
Exploration of host-microbial relationships and transmission within an intensively managed
population is required to understand the epidemiology of infectious organisms in fragmented
populations, and to assess the disease risks associated with the fragmentation and translocation of
endangered wildlife. Modelling the prevalence of commensal Campylobacter spp. within highly
managed takahe populations, indicates that anthropogenic management of endangered species within
altered environments may have unforeseen effects on microbial exposure and carriage. The results of
this study suggest that the transmission dynamics and carriage of bacterial species is heterogeneous
between populations. The prevalence of enteric Campylobacter spp. in takahe differed according to a
number of population and environmental variables. C. sp. nova 1 showed strong host-commensal
relationships, whilst the prevalence of C. jejuni and C. coli were associated with variables relating to
age, management and environmental exposures of the host. Whilst modelling of a commensal may not
directly simulate the dynamics of a pathogen within fragmented populations due to the lack of
pathogenicity and recovery of the host, it has been a good proxy to explore how anthropogenic
changes in host population dynamics can influence microbial carriage. The consequences of
management-associated influences on host-microbial relationships observed in this study are difficult
to predict, but may result in increased exposure to pathogens of significance to wildlife health.
Forced perturbation of populations through frequent translocation of takahe between the
geographically isolated locations (Grange et al. 2014), for reasons of conservation and inbreeding
(Jamieson et al. 2006) provides a means of population-wide mixing of birds where subpopulation
barriers are bypassed. The high prevalence of C. sp. nova 1 in all takahe populations supports the
hypothesis of dynamic mixing of infectious organisms irrespective of sub-population biotic and
abiotic factors. Despite extensive examination of multiple hosts, including farmed livestock and
wildlife, C. sp. nova 1 has only been identified in New Zealand surface water and members of the
Rallidae family (French et al. 2014). Extant takahe originate from a small population in the
Murchison Mountains (Ballance 2001). If C. sp. nova 1 has had a long history of takahe colonisation,
niche adaptations to the host may have allowed survival, persistence and transmission of the species
through the generations, resulting in the current high prevalence in extant populations. C. sp. nova 1 is
newly described, and while the identification methods used in this investigation are unable to observe
64
differences between isolates, future in depth analyses may identify differences between these isolates
(Chapter 4).
Differential carriage of C. coli and C. jejuni between takahe sub-populations supports the theoretical
hypothesis that the takahe sub-populations would vary in their propensity to maintain and transmit
infectious organisms (Grange et al. 2014). Depending on the study species and ecosystem dynamics
involved, translocated or artificially reared animals may be at an advantage in terms of host parasitism
compared to their home range. The parasite release hypothesis suggests that animals may escape
native pathogens when moved to a new location (Torchin et al. 2003). Reintroduced native species are
selected for good health prior to transfer, consequently disease circulating in this select group may be
lower than the source population (Almberg et al. 2012). Many pathogens have complex life cycles
requiring more than one host and thus if this host is not present in the new location, the pathogen may
not be maintained. Translocated populations of takahe may have escaped from micro-organisms
present in the founding population, through random selection of microbe free individuals or the
absence of reservoirs in the new environment. Either hypothesis could explain the higher prevalence
of C. jejuni in the wild Murchison Mountain population.
Translocated populations are expected to acquire pathogens circulating within the environment
(Almberg et al. 2012), but it is thought to occur at a slower rate than what they would be exposed to in
their original geographic range (Torchin et al. 2003; Torchin & Mitchell 2004). Translocated
populations could acquire novel pathogens via spill-over from alternative hosts within the new
environment (Tompkins et al. 2011). Contrary to the traditional dogma, domestic animals living in
close proximity to wildlife could be a reservoir of exotic pathogens to native species, potentially
contributing to their decline (Besser et al. 2013; Besser et al. 2012a; Besser et al. 2012b).
C. coli is frequently isolated from faecal material derived from avian populations whom exhibit no
obvious clinical signs of disease (Benskin et al. 2009). Many C. coli sequence types are
epidemiologically associated with livestock, including pigs and chickens (Kittl et al. 2013).
Campylobacter spp. are able to survive in the environment (Bull et al. 2006), thus livestock-associated
C. coli may have been transmitted from livestock to takahe populations in areas associated with
farmland, via faecal contamination of shared resources and/or other wildlife passing between farms
and takahe locations. This transmission theory could also explain the higher diversity of
65
Campylobacter spp. associated with takahe born in populations close to farmland. However, carriage
of multiple species is not necessarily an indication of being a persistent host to multiple species, as
some infections may be transient due to temporal patterns of shedding (Colles et al. 2009).
Early exposure appears associated with microbial carriage in takahe, with a higher prevalence of C.
coli reported in juveniles than adults. Similar findings were observed in a study of C. jejuni in
European starlings (Sturnus vulgaris) where shedding of the bacteria was higher in young birds
(Colles et al. 2009). Carriage of C. coli in juveniles is unlikely to be attributable to differences in
behaviour, as juvenile takahe shadow adults for up to two years prior to independence (Maxwell &
Jamieson 1997). Juvenile takahe may be more susceptible to infection due to the lack of acquired
immunity. However, it has been recognised that further investigations are required to understand the
role immunity may play in carriage of Campylobacter spp. in wild birds (Waldenstrom & Griekspoor
2014). In accordance with other studies of Campylobacter spp. in wild avian populations (Colles et al.
2009; Waldenstrom et al. 2002), sex was not significantly associated with carriage of Campylobacter
spp. in takahe. A common outcome of multivariate modelling in this study was the involvement of
nest environment in explaining the carriage of Campylobacter in takahe. Although the mechanisms by
which early rearing environment affect the gastrointestinal microbiome of takahe is unclear, there
appears to be an association between nest location attributes and the carriage of Campylobacter spp.
later in life. In the early days of takahe recovery management, cross-fostering was used where a
pukeko (Porphyrio porphyrio) foster mother was trained to rear foster takahe chicks (Bunin &
Jamieson 1996). Although the two species are closely related, this practice could have unwittingly
introduced exotic micro-organisms to the takahe population. However, the rearing method of an
individual was not significantly associated with the carriage of Campylobacter spp.
Although Campylobacter spp. are not thought to be pathogenic to takahe, exposure to non-native
Campylobacter spp. may result in reduction of fitness (Waldenstrom et al. 2010). Several unidentified
Campylobacter spp. were isolated from takahe, the pathogenicity of which remains unknown.
Translocated individuals and isolated populations with few immigrants are potentially more likely to
carry unidentified Campylobacter spp. Two hypotheses could explain this association; translocated
populations may have gained Campylobacter spp. from novel reservoirs, and / or Campylobacter spp.
may have evolved following isolation of takahe.
66
Limitations in the interpretation of the findings of this study are evident due to the limited sample
size, missing information, imperfect tests and potential confounding variables. For example, the
breeding and wild subpopulations were comprised of single locations, whereas the insurance group
includes many sites. Therefore, populations may be confounded by location specific attributes which
were beyond the scope of the study. It is likely that a combination of biotic and abiotic factors
explains the observed carriage of different species of Campylobacter in takahe in this study. Temporal
and seasonal trends in the carriage of Campylobacter spp. in wild birds have been reported
(Waldenstrom & Griekspoor 2014). Although this study found some associations with the time period
samples were collected, the opportunistic design and confounding variables, such as sampling certain
locations at a certain time of year, prevents a true assessment of temporal trends in takahe.
Intensive conservation management of endangered species in fragmented populations can influence
host-microbial relationships. The disease risk posed to translocated populations of takahe may be
determined by the choice of habitat in which they are maintained. Implementing buffer zones around
reserves is recommended to reduce the risk of pathogen transmission from reservoir species to
endangered wildlife. Additionally, relatively unmanaged populations appear to have different
microbial carriage compared to captive populations. Therefore, the movement of animals between
these populations may transmit pathogens to naive individuals, and thus would not be advised. This
study has provided a good basis for further investigation of microbial dynamics in translocated
populations and has the potential to inform risk assessments and aid in our understanding of the
epidemiology of infectious disease in wildlife.
3.6. Acknowledgements
This study was funded by the Allan Wilson Centre. Samples were collected under a Massey
University animal ethics permit MUAEC Protocol 11/95. We would like to thank S. Carver, A. S.
Browne, S. Michael, P. Marsh, G. Greaves, A. Wilson, B. Jackson and the Friends of Tiritiri Matangi
for assistance and the Department of Conservation and the Maori community for their support. We
would like to acknowledge the reviewers whom provided useful insights into this manuscript.
67
3.7. Supporting information
R script for LCA in diagnostic testing (Appendix 9.3-1), detailed LCA model outputs (Appendix
9.3-2), MCA plots (Appendix 9.3-3), and univariate Campylobacter spp. model results (Appendix
9.3-4, 5, 6, 7, 8) are available in the appendix.
68
CHAPTER 4
WILDLIFE TRANSLOCATION AND THE EVOLUTION AND
POPULATION STRUCTURE OF A HOST-ASSOCIATED
COMMENSAL CAMPYLOBACTER SPP.
69
70
4. Wildlife translocation and the evolution and population structure of a
host-associated commensal Campylobacter spp.
4.1. Abstract
There is an increasing need for the conservation management of threatened wildlife, and yet we have
a limited understanding of the effects of tools such as translocation has upon pathogen transmission
and disease ecology. Our ability to predict and control outbreaks of infectious disease is hampered by
the complex interactions between hosts, pathogens and the environment. The use of commensal
bacteria as a proxy for invasive pathogens can help our understanding of some of the epidemiological
features of infectious diseases, such as microbial transmission and evolution in natural systems. The
genomic structure of a prevalent rail-associated endemic bacterium, Campylobacter sp. nova 1, was
explored in a well-described population of endangered takahe (Porphyrio hochstetteri). The
distinctive population structure of translocated takahe provides a unique opportunity to investigate the
influence of host isolation on enteric microbial diversity. Whole genome sequencing and ribosomal
multi-locus sequence typing (rMLST) was carried out on C. sp. nova 1 isolated from one third of the
takahe population from multiple locations. C. sp. nova 1 was genomically diverse and multivariate
analysis of 52 rMLST alleles revealed location-associated differentiation of C. sp. nova 1 sequence
types. While there was evidence of recombination between lineages, bacterial divergence appears to
have occurred in the face of host range expansion and isolation. Anthropogenic management of
wildlife in highly connected locations such as a breeding centre may create artificial opportunities for
exposure and mixing of a broad selection of genotypes. These results support the idea that restricted
gene flow in fragmented populations affects the genomic differentiation of microbes. Subtle but
important differences in host-microbe relationships as a consequence of management may result in the
emergence of pathogens and have important implications when relocating allopatric wildlife
populations.
71
4.2. Introduction
Emerging infectious diseases (EIDs) in wildlife are of increasing concern due to their direct impact on
biodiversity and the risk of zoonotic transmission (Daszak et al. 2000; Jones et al. 2008), with over
half (54%) of human EID events attributable to bacterial pathogens of zoonotic origin (Jones et al.
2008). Effective management and mitigation of infectious wildlife disease requires an understanding
of both biotic and abiotic factors that drive pathogen transmission and persistence within ecosystems.
Although infectious disease is rarely the sole driver of extinction (Heard et al. 2013; Smith et al.
2006), it may contribute to local population declines (Berger et al. 1998; Fisher et al. 2012), affect the
success of reintroduction programmes (Almberg et al. 2012) and thus the prospect of long term
species survival. Our understanding of the epidemiology of infectious diseases in wildlife is often
limited by the complex interactions between hosts, pathogens and the environment, and difficulties
often encountered when surveying wildlife populations.
Epidemics of invasive pathogens have had significant impacts on wildlife. The spread of chytrid
fungus (Batrachochytrium dendrobatidis) has been associated with high host mortality or morbidity in
amphibians worldwide (Rachowicz et al. 2006; Skerratt et al. 2007). Studies of host-pathogen
relationships during an epidemic require an in situ outbreak situation. Not only is this a rare
occurrence but there are several logistic difficulties associated with studying uncontrolled and
unknown pathogens, as exemplified by the 2014 pandemic of Ebola virus in West Africa
(Kupferschmidt 2014). Endemic pathogens circulating in a population may regulate host populations
but could be more difficult to detect and characterise. An alternative is to study the transmission and
population biology of a common host-associated commensal in natural ecosystems as a proxy for an
epidemic pathogen (Bull et al. 2012; Chapter 3 ; Chiyo et al. 2014). Such an approach allows key
epidemiological and ecological features, such as microbial transmission and evolution to be studied in
the absence of disease. Additionally, a commensal in one host may be a pathogen in another
(Waldenstrom et al. 2010), which has implications when reconnecting allopatric species through
wildlife conservation practices (Cunningham 1996; Daszak et al. 2000).
Heterogeneous social (Bull et al. 2012; Porphyre et al. 2011; VanderWaal et al. 2013b) and meta-
population dynamics (Cowled et al. 2012; Guivier et al. 2011) present in natural populations have
72
been shown to affect microbial diversity and transmission of infectious agents. However, much of our
knowledge is based upon theoretical evidence. Robust empirical epidemiological studies of pathogen
dynamics in multiple populations require a high prevalence, strong host association and a well-
described host population structure. Endangered wild animals are often maintained in relatively small
and fragmented sub-populations within highly altered environments. These well-documented
populations provide ideal model systems for examining factors contributing to the transmission and
population biology of microorganisms including enteric pathogens and commensals.
The model system under investigation in this study captures the carriage, transmission, population
structure and evolution of a prevalent, host-associated enteric Campylobacter species nova 1 in takahe
(Porphyrio hochstetteri). Takahe are an endangered flightless rail endemic to New Zealand (BirdLife
International 2013) with a population of approximately 230 individuals (Wickes et al. 2009). The
historic translocation of birds from the one remaining wild population in the Murchison Mountains,
New Zealand (Ballance 2001) to multiple wildlife sanctuaries has created a network of small sub-
populations (Grange et al. 2014). Although fragmented, takahe populations are connected through
well-described translocation networks, allowing the effects of within and between microbial
population dynamics to be considered (Grange et al. 2014). Social network models have indicated the
presence of a dynamic network of sink, source and hub populations with implications for pathogen
carriage, transmission and evolution (Grange et al. 2014).
A prior investigation showed that takahe sub-population attributes including management and
proximity to agriculture influenced faecal carriage of commensal Campylobacter jejuni and
Campylobacter coli respectively (Chapter 3). The study examined prevalence at the Campylobacter
species level but lacked the resolution required to explore the microbial population structure of more
prevalent Campylobacter spp. such as Campylobacter species nova 1, a proposed novel
Campylobacter spp. (French et al. 2014) which was isolated from 90% of takahe tested (Chapter 3).
In the first study of its kind, high resolution genotyping was used to model the effects of host
population history and spatial characteristics of the environment on the population structure of an
enteric microorganism in a well-characterised host population.
73
4.3. Methods
4.3.1. Sample collection and culture
Cloacal swabs were opportunistically collected from takahe resident in 6 locations in New Zealand
(Figure 4-1) during pre-translocation disease screening or annual health checks between November
2011 and April 2013. Samples were stored in Aimes Charcoal transport media (Copan, California,
USA) and refrigerated at 4oC for up to 7 days prior to laboratory processing. Previously described
protocols were used for microbial culture and extraction of Campylobacter spp. DNA for PCR
identification of C. sp. nova 1 positive isolates (Chapter 3). This PCR targets a short section of a
putative C4-dicarboxylate trans-membrane transport gene believed to be found only in C. sp. nova 1
(Chapter 3).
4.3.2. Selection of C. sp. nova 1 for genomic sequencing
A selection of 70 C. sp. nova 1 isolates were chosen for comparative whole genome sequencing.
These isolates originated from 69 (plus 1 repeat) takahe spread between six locations in New Zealand:
Burwood Bush breeding centre (n = 22), Murchison Mountains (n = 25), Mana Island (n = 1), Maud
Island (n = 4), Maungatautari reserve (n = 4) and a private island reserve which is undisclosed due to
confidentiality restrictions (n = 14) (Figure 4-1, Appendix 9.4-1). One takahe moved from
Maungatautari reserve to Burwood Bush during the study and isolates from both locations were
analysed.
4.3.3. Genomic DNA preparation and processing
Pure cultures were recovered from glycerol broth stored at -80oC and grown in micro-aerobic
conditions at 42oC on Columbia Horse Blood agar (Fort Richard, Auckland, New Zealand). DNA
was extracted using the Qiagen DNeasy blood and tissue kit (Bio-Strategy, Auckland, New Zealand).
Protocols were followed according to the manufacturer’s instructions for gram negative bacteria with
the final elution step modified to 200µl sterile milliQ water (Merck, Palmerston North, New Zealand).
DNA was checked for quality using Qubit dsDNA HS/RNA/Protein assay kits (Life Technologies,
Auckland, New Zealand), and stored at -20 oC prior to sequencing at the New Zealand Genomics Ltd.
74
Genomes were sequenced using an Illumina MiSeq instrument (Illumina, Victoria, Australia)
according to the manufacturer’s instructions with paired read lengths each of 250 base pairs. DNA
samples were fragmented by nebulisation for 6 minutes at a pressure of 32 psi, purified, then end
repaired, A-tailed, adaptor-ligated, fractionated, purified and enriched according to the manufacturer's
instructions, using the TruSeq DNA LT Sample Prep Kit v2-Set A and B (Illumina, Victoria,
Australia). The prepared libraries were normalized to equal molarity, diluted to 2nM and pooled in
libraries of 20 samples. A flow cell was prepared for each pool and sequencing reactions using 9
pmoles of the pooled libraries were performed with the MiSeq Reagent Kit v2 (Illumina, Victoria,
Australia) to give approximately 12 to 15 million clusters per run.
4.3.4. Genome assembly, curation and annotation
The algorithm package Velvet (version 1.2.10) (Zerbino & Birney 2008) was used for de novo
genome assembly and alignment. First, short reads sequences of 250 base pairs were broken into
smaller sequences (k-mers). Then, a de Bruijn graph was constructed from those short sequences. The
sequences were assembled across a range of k-mer lengths in increments of 10 between 215 and 65
base pairs. The resulting Velvet contiguous sequences (contigs) were stored in a MySQL (version 5.6)
database. The best assembly per genome was chosen using a Perl-based in-house ranking system
using a score derived from the number of contigs, size of contigs, assembly length and the N50 score.
Concatenated contigs within an assembly were annotated with the program Prokka (version 1.9)
(Seemann 2014). Predicted genes from the annotated genomes were then clustered using OrthoMCL
(version 2.09) (Li et al. 2003) and subsequent in-house parsing.
4.3.5. Ribosomal multi locus sequence typing (rMLST) of C. sp. nova 1
Nucleotide sequences for 52 of the 53 genes (the order Campylobacterales does not possess the rpmD
gene) encoding bacterial ribosomal protein subunits (rps) used for rMLST (Jolley et al. 2012), were
identified and extracted from the assembled genomes primarily on annotation, but also using a C.
jejuni reference set of rMLST genes as a database for a BLAST search. Using custom Perl scripts,
unique alleles for each gene were defined based on their nucleotide sequence. Each genome was
assigned a custom sequence type number defined by the allelic profile of the 52 gene numerical
75
combination. Each individual gene was aligned using Muscle (Edgar 2004), and these were
concatenated to make a single alignment per genome. Both allelic profile data and concatenated
rMLST gene sequences were used in subsequent analyses.
Three distance matrices were created from pairwise comparison of 52 gene rMLST profiles of the 70
C. sp, nova 1 genomes (Appendix 9.4-1). A haplotype distance matrix was constructed from the 52
rMLST allelic dataset using the GenAIEx Microsoft Excel add on package (Peakall & Smouse 2012).
Two pairwise nucleotide distance matrices were created using uncorrected P measure and general
time reversible (GTR) in SplitsTree4 (Huson & Bryant 2006). The uncorrected P measure was
obtained by dividing the number of single nucleotide differences by the total number of nucleotides
compared between two gene sequences. The GTR model accounts for nucleotide substitution where it
assumes a symmetric substitution between bases, for example C change into G at the same rate that C
changes into C. Additionally, the GTR model allows for input of variable base frequencies. Values of:
A: 35%, C: 15%, G: 15% and T: 35%, were used as this was a good approximation for the GC content
of C. sp. nova 1.
4.3.6. Core genome and rMLST tree construction
A selection of publicly available Campylobacter spp. whole genome sequences were downloaded
from the NCBI database (Benson et al. 2009; Sayers et al. 2009) (Appendix 9.4-2) and were subject to
the methods previously described for the extraction and processing of rMLST sequences.
NeighborNet trees of the published Campylobacter spp. rMLST sequences and the 70 C. sp. nova 1
isolates were visualised in SplitsTree4 (Huson & Bryant 2006). The tree was constructed from a
pairwise comparison of genomic distance using uncorrected P measures (Appendix 9.4-1).
Amino acid sequences of orthologous genes, those which were inherited vertically from a common
ancestor, were used for core genome comparison of the 70 C. sp. nova 1 isolates. This was
implemented using a custom Perl script and clustered using OrthoMCL version 2.0.9 (Li et al. 2003).
A NeighborNet tree was created from a distance matrix of uncorrected P distances based on pairwise
comparison of core gene amino acid sequences of the same length, excluding genes of different
length. The tree was then visualised in SplitsTree4 (Huson & Bryant 2006).
76
4.3.7. Multivariate analysis of the relationship between location factors and genetic distance
The three matrices were explored using multidimensional scaling (MDS) followed by permutational
multivariate analysis of variance (PERMANOVA) within Primer v6 software (Clarke 1993). Factors
used in the PERMANOVA model included sample site and takahe nest site, as well as co-factors
relating to location geography (North vs South Island, Island vs Mainland) and management (Wild vs
Captive breeding and insurance, Captive breeding vs Insurance). The term insurance is defined as
sub-populations maintained in free to roam reserves external to the founding wild population.
PERMANOVA models were created with 9,999 permutations using a method which allows
unrestricted permutation of raw data. Non-significant factors and interactions were removed in a
stepwise manner until the final model containing significant factors and corresponding cofactors
remained.
Overall and pairwise genetic distances between C. sp. nova 1 allelic profiles were examined in
relation to significant factors identified from the PERMANOVA analysis. Fixation index (FST) values,
a measure of population differentiation due to genetic structure, were calculated using analysis of
molecular variance (AMOVA) methods within Arlequin software (Excoffier & Lischer 2010).
Pairwise FST values were visualised in SplitsTree4 (Huson & Bryant 2006).
4.4. Results
4.4.1. C. sp. nova 1 comparison to published Campylobacter spp.
All 70 C. sp. nova 1 isolates formed a distinct cluster, separate from other published Campylobacter
spp. included in this study (Figure 4-2). rMLST gene comparison of the Campylobacter spp. genomes
correlates with previous investigations (French et al. 2014) where C. sp. nova 1 was closely related to
Campylobacter coli and Campylobacter jejuni but was genomically distinct.
77
Figure 4-1 Map of takahe (Porphyrio hochstetteri) sampling locations in New Zealand. The private island location is undisclosed due to confidentiality restrictions. Pie charts represent the number and clade types, see Figure 4-3, of Campylobacter sp. nova 1 isolated from each location.
78
Figure 4-2 NeighborNet tree based on pairwise comparison of rMLST nucleotide sequences of Campylobacter sp. nova 1 isolated from takahe (Porphyrio hochstetteri) and a selection of publicly available Campylobacter spp.
79
Figure 4-3 NeighborNet trees of 70 Campylobacter sp. nova 1 using unmeasured P distances of a) amino acid core genome comparisons and b) nucleotide comparison of 52 rMLST genes. Dots represent individual C. sp. nova 1 genomes isolated from takahe (Porphyrio hochstetteri), coloured by location. Numbers were assigned to each distinct clade on a branch of the tree.
80
4.4.2. Genomic differentiation of C. sp. nova 1 isolates
Visualisation of nucleotide differences between isolates of C. sp. nova 1 isolates using pairwise
uncorrected P distances of the 52 rMLST genes and core genome illustrates fine scale genomic
diversity among the isolates (Figure 4-3). Seven clades were observed using both methods, with
evidence of genetic recombination between isolates revealed by the net-like structure in the centre of
the star shaped tree. These structures represent uncertainty of the rMLST gene ancestry which is
characteristic of recombination. The positions of isolates on the tree suggested a correlation between
clade and the location of the host from which the bacterium was isolated (Figure 4-3). However,
sequence types isolated from Burwood Bush breeding centre are genomically diverse with
representatives from four different clades being isolated from takahe within this location.
4.4.3. Multivariate analysis of C. sp. nova 1 rMLST allelic profiles
Initial exploration of C. sp. nova 1 rMLST allelic profiles using multidimensional scaling (MDS)
(Appendix 9.4-3) identified three clusters according to similarity in their allelic profile. Large
distances were evident between clusters, with some defined grouping of isolates structured within a
cluster, particularly within cluster two. Clusters one and three correlated well to clades one and two
defined within the NeighborNet tree, whilst MDS cluster number 2 was comprised of all remaining
NeighborNet clades (Appendix 9.4-3, Figure 4-3).
Variables relating to sample and host nest location, including co-factors defined by location
geography and management, were explored in association with rMLST allelic profiles (Table 4-1),
unmeasured P (Appendix 9.4-4) and GTR distance matrices (Appendix 9.4-5) using PERMANOVA,
with final models containing only significant factors. All three methods provided similar results with
location being a significant factor (P(perm) = 0.0001) explaining the genetic distance observed in
pairwise comparisons of rMLST profiles. Genomic differentiation of C. sp. nova 1 according to
location may be attributable to co-factors relating to management or historic isolation, with C. sp.
nova 1 isolated from the wild being different from those in the breeding and insurance populations
(P(perm) = 0.001) (Table 4-1). The sample size did not provide sufficient statistical power to explore
interaction terms between sampling location and nest site. Additionally there was no significant
81
contribution of nest location or its corresponding co-factors on the observed genomic structuring of C.
sp. nova 1.
Factors d.f SS Est. of variance Pseudo-F P(perm) values Perm
Sampling location 5 9859.9 142.67 3.792 0.0001 9913
Cofactors
North vs South Island 1 1101.7 52.062 1.7819 0.1211 9149
Wild vs Captive breeding & Insurance 1 3204.4 81.42 5.4559 0.0014 9814
Captive breeding vs Insurance 1 1714.1 45.236 2.46 0.0486 9938
Island vs Mainland 1 1520.1 32.798 2.4835 0.0525 9953
Residuals 64 33283 520.04 Total 69 43143 Table 4-1 Reduced PERMANOVA model from 9,999 permutations (perm) containing significant factors and cofactors
associated with rMLST allelic profiles of Campylobacter sp. nova 1. The cofactor insurance is defined as sub-populations maintained in free to roam reserves external to the founding wild population.
Support for strong genetic differentiation of populations according to location was observed with an
FST value of 0.22 (P < 0.01), where most variation was within location defined populations (77.8%,
d.f. 64, SS 822.75), with the remaining explained among populations (22.2%, d.f. 5, SS 251.37).
Pairwise comparison of location FST values identified some heterogeneity in the genetic distance of
isolate populations between locations (Figure 4-4). Branch length and position on the tree correlate
with genetic distance. The star-like structure and long branch lengths would suggest location was
strongly associated with bacterial divergence. Notably, Burwood Bush breeding centre was placed in
the centre of the tree with no apparent branch. Thus isolates derived from this location were less
distinct from isolates in other locations included in the analysis, compared to all other pairwise
comparisons.
82
Figure 4-4 FST tree based on pairwise comparison of 70 rMLST profiles of Campylobacter sp. nova 1 isolated from takahe (Porphyrio hochstetteri) populations.
83
4.5. Discussion
Determining the impacts of host meta-population dynamics on bacterial spread and maintenance in a
population has been a key focus of epidemiological studies. The emergence of robust methods
integrating survey and genomic approaches allow complex analyses of the relationships between host-
associated variables and bacterial genotype (Biek et al. 2012; Girard et al. 2004). This study provides
evidence of spatial clustering of bacterial genotypes of a commensal bacterium, C. sp. nova 1, carried
by endangered populations of takahe. Microbial genotypes and rMLST allelic profiles were strongly
associated with geographic location, with population isolation and environmental factors potentially
contributing to the observed diversity of C. sp. nova 1 sequence types isolated from its host. Our
results indicate that fragmentation and anthropogenic manipulation of populations may influence host-
microbial relationships, with potential implications on niche adaptation and evolution of microbes in
remote environments. Studies of this nature lend insight into the epidemiology of pathogens in natural
systems which may threaten wildlife populations or become significant zoonoses.
Environmental attributes can restrict or enhance host population structure. This is particularly
pertinent with respect to wildlife whose connectivity between populations has been adversely affected
by habitat loss and fragmentation (Hand et al. 2014). Reduced mixing of hosts could have
downstream effects on the patterns of disease spread, pathogen population structure and gene-
exchange (Gandon et al. 2008). Takahe populations underwent a significant population bottleneck and
range contraction to one isolated mountain population due to anthropogenic impacts (Bunin &
Jamieson 1995; Trewick & Worthy 2001). C. sp. nova 1 is thought to be endemic to takahe with
prevalence estimates of 90% (Chapter 3). A host adapted ancestral lineage of the bacterium may have
been present in the founding Murchison Mountain population which subsequently evolved
biogeographically as a result of host range expansion, and distribution of populations into multiple
new and isolated locations (Ballance 2001). This may explain the observed heterogeneity and spatial
clustering of C. sp. nova 1 isolated from takahe in this study, as well as the genomic separation of
genotypes isolated from takahe in the Murchison Mountains from those in insurance populations
(Figure 4-5A). An alternative hypothesis is that the ancestral population in the Murchison Mountains
84
consisted of a diverse range of genotypes, bottlenecking then occurred through the translocation of
takahe, resulting in a random subset in each population (Figure 4-5B).
A variety of host attributes and behaviour may influence host exposure, prevalence and transmission
of microbes. The complexity of such interactions have been explored using social network scientific
methods in an epidemiological context (Bull et al. 2012; Chiyo et al. 2014; Hirsch et al. 2013;
Porphyre et al. 2011; Rushmore et al. 2013; VanderWaal et al. 2013a). The analysis of population
connectivity with respect to the carriage of C. sp. nova 1 genotypes was not able to be conducted in
this study. This was due to the fact that each takahe population used in this study possessed unique
network measures derived from historic takahe movements (Chapter 2; Grange et al. 2014).
Social interactions between hosts may be more influential on sharing of genotypes than spatial
proximity of hosts within an environment (Bull et al. 2012; VanderWaal et al. 2013b). Immigration
and emigration of takahe between sub-populations is controlled and mediated by human intervention
(Wickes et al. 2009). Therefore, there remain opportunities for continued bacterial gene flow between
supposedly allopatric populations. In theory, this would have a homogenising effect on sequence type
carriage either through competitive exclusion where dominant sequence types out-compete extant
microbial flora, or via genomic recombination and stabilising selection. Carriage of panmictic
Escherichia coli genotypes isolated from social groups of elephants (Loxodonta africana) sharing
sources provided support for this concept (Chiyo et al. 2014). However, this study of C sp. nova 1 in
takahe provides evidence that bacterial divergence could occur in the face of population isolation with
continued gene flow and within relatively short time frames. Initial studies by French et al. (2014)
investigated the phylogeny of C. sp. nova 1 isolated from New Zealand water and members of the
Rallidae family. Analysis suggested two clades of the species diverged over 1000 years ago (French et
al. 2014). If the divergence of C. sp. nova 1 clades isolated from takahe in this study occurred within
in similar time frame, it seems unlikely that evolution following translocation from the Murchison
Mountain population in the last 30 years (Bunin et al. 1997) explains the biogeographic clustering
observed in extant takahe populations.
85
Figure 4-5 Schematic of hypotheses to explain geographic clustering of Campylobacter sp. nova 1 genotypes observed in the takahe (Porphyrio hochstetteri) sup-populations. Coloured dots represent C. sp. nova 1 genotypes. (A) By the allopatric speciation hypothesis, there was an ancestral population of C. sp. nova 1 in the Murchison Mountains, and then following translocation and isolation these diverge as a result of rapid genetic drift. (B) By the selection hypothesis, the ancestral population contained a diverse range of genotypes, then bottlenecking occurred through translocation of takahe, with a random subset in each population. (C) By the connectivity hypothesis, the observed biogeographic diversity of genotypes is the result of population connectivity, with ‘hub’ populations containing multiple genotypes and more isolated populations containing few genotypes due to isolation. (D) By the reservoir hypothesis, the population C. sp. nova 1 genotypes present within a population were determined by the presence and transmission of bacteria from environmental and alternative hosts. Alternatively, the observed diversity is a result of any combination of A to D.
86
Sequence analysis of C. sp. nova 1 revealed a diverse range of genotypes, with location confounded
attributes explaining most of the genetic differentiation. This is suggestive of rapid genomic
divergence as a consequence of takahe isolation and is in agreement with previous network and
molecular investigations (Chapter 3 ; Grange et al. 2014). Notably, the breeding centre population
contained a diverse range of genotypes, with representatives from multiple clusters. This site acts as a
hub of the population network with regular immigration from multiple sources (Grange et al. 2014).
Unlike most other populations tested where takahe are free to roam and mix socially, takahe within
the breeding centre are separated into captive pens according to family groups. This management
practice may prevent mixing and allow the maintenance of a diverse range of C. sp. nova 1 genotypes
in the breeding centre population which have been derived from other locations, as observed in this
study (Figure 4-5C).
Evolution of a microbe within allopatric populations is a concern when considering the management
of endangered species. A commensal in one host may be a pathogen in another, and thus the
translocation of an infected individual into a naive population may result in morbidity and mortality in
a new host. An infection experiment conducted by Waldenstrom et al. (2010), demonstrated a marked
decrease in body mass of European robins (Erithacus rubecula) following inoculation with a C. jejuni
isolate derived from another avian species. Similar spatial clustering of microbial genotypes has been
observed in the spread of a pathogen in wild populations. Rapid spread of a single multi-locus
Yersinia pestis genotype within a prairie dog (Cynomys ludovicianus) population was followed by
localised microbial differentiation within prairie dog sub-populations, which may be attributable to
isolation or local maintenance in cryptic mammalian reservoirs (Girard et al. 2004). Modelling of
canine distemper virus outbreaks in the Serengeti, suggests an important role of local reservoir species
as a source of in pathogen spill-over to African lions (Panthera leo) (Craft et al. 2009). It is
reasonable to hypothesise that the observed spatial clustering of C. sp. nova 1 sequence types may be
a result of local interactions with reservoir species (Figure 4-5D). C. sp. nova 1 has previously been
isolated from New Zealand surface water and members of the Rallidae family (French et al. 2014).
Given the evidence of strong host-association (Chapter 3) and genomic complexity of C. sp. nova 1
which was previously unobserved from traditional PCR typing methods (Chapter 3), it seems likely
87
that a combination of interactions between resident takahe, translocation and co-infections of hosts
may contribute more to local genomic differentiation (Figure 4-5).
Epidemiological data garnered from investigations of highly managed takahe highlights the need for
risk analysis and pathogen screening when considering movements of individuals between locations.
Insight gained from this study not only has implications for the management of takahe; it also applies
to the relocation of any animal including domestic livestock. Complex agricultural networks regularly
move livestock across natural barriers and thus have a high risk of pathogen transmission resulting in
outbreaks, as exemplified in the dissemination of foot and mouth disease virus between farms in the
United Kingdom 2001 (Ortiz-Pelaez et al. 2006). Gaining a full understanding of pathogen
epidemiology in wildlife systems is a challenging prospect. In this study, the small sample size and
selection of a single isolate from an individual host limits our interpretation of location associations.
For example, the methods used do not account for carriage of multiple C. sp. nova 1 genotypes within
a single host. This is a likely occurrence given the evidence of multiple carriage of Campylobacter
spp. in takahe (Chapter 3). The use of an endemic commensal as a proxy for invasive pathogens can
help our understanding of some of the epidemiological features of infectious disease, such as
transmission and evolution, but this approach will always be limited by the lack of pathogenicity or
recovery from infection. However, with the advent of whole genome analyses which can be applied
across multiple bacterial species (Jolley et al. 2012) and robust phylogenetic methods for the
exploration of factors that influence host-microbe relationships, these methods may lead the way in
understanding and mitigating disease risks imposed by the anthropogenic management of wildlife.
4.6. Acknowledgements
This study was funded by the Allan Wilson Centre. Samples were collected under a Massey
University animal ethics permit MUAEC Protocol 11/95. I would like to thank A. S. Browne, P.
Marsh, G. Greaves, A. Wilson, B. Jackson and the Friends of Tiritiri Matangi for assistance and the
Department of Conservation and the Maori community for their support.
88
4.7. Supporting information
Appendix 9.4-1 is available electronically on a disc at the end of this thesis. The list of publically
available Campylobacter spp. genomes used (Appendix 9.4-2), the MDS plot (Appendix 9.4-3) and
PERMANOVA results using the uncorrected P measure (Appendix 9.4-4) and GTR matrix (Appendix
9.4-5) are available in the appendix.
89
90
CHAPTER 5
INVESTIGATION OF VERTEBRATE RESERVOIRS OF
CAMPYLOBACTER SPP. IN AN ISLAND ECOSYSTEM
91
92
5. Investigation of vertebrate reservoirs of Campylobacter spp. in an island
ecosystem
5.1. Abstract
Prediction and prevention of infectious disease in wildlife populations requires a baseline understanding of
pathogen transmission, often involving multiple host species in complex ecosystems. Identifying reservoir
populations which maintain and transmit pathogens to endangered species is a challenging but important area
of epidemiological investigation. In order to explore the evolution and transmission of bacteria between hosts
and reservoirs, this study determined the prevalence and genomic composition of enteric Campylobacter spp.
in five vertebrate orders occupying an island ecosystem used for conservation of endangered takahe
(Porphyrio hochstetteri). The genus Campylobacter contains a number of important pathogens of humans,
domestic animals and wildlife but many are considered harmless commensal bacteria in certain hosts.
Culture positive Campylobacter spp. isolates were subject to PCR for identification, followed by whole
genome analysis. Prevalence of Campylobacter spp. was variable among sympatric vertebrate communities
with estimates ranging from 0% to 100%. Ribosomal multi-locus sequence typing (rMLST) analysis of
nineteen Campylobacter samples isolated from passerines and rails provided evidence of host-associated but
not exclusively host-specific Campylobacter spp. genotypes. Genetic similarities were observed between
Campylobacter spp. genomes isolated within a host species and between closely related hosts. Evidence of
allele sharing suggests a common evolutionary history and potential recombination between species of
Campylobacter colonising different host species. Exploring host-microbe relationships through the use of
genomic sequencing of commensal bacteria in a multi-host system study provides insight into potential
epidemiological interactions between hosts and pathogens, including the roles that sympatric communities
may play in epidemics of more pathogenic generalist microorganisms.
93
5.2. Introduction
Predicting the emergence of wildlife disease and developing mitigation strategies requires an understanding
of complex heterogeneous ecological interactions and the sharing of pathogens between communities. The
recent integration of genomics, bioinformatics, epidemiology and ecology provides a combination of tools in
which the ecological and evolutionary linkages between host and pathogen may be deconstructed and inter-
host transmission dynamics inferred. Reservoirs are regarded as populations which maintain and transmit a
pathogen to epidemiologically connected vulnerable target populations, such as endangered species (Haydon
et al. 2002). It is often challenging to identify reservoirs of infectious organisms in susceptible hosts. For
instance, evidence indicates interspecific interactions between African lions (Panthera leo), spotted hyenas
(Crocuta crocuta) and jackals (Canis spp.) contribute to outbreaks of canine distemper virus within lion
populations in the Serengeti (Craft et al. 2008; Craft et al. 2009).
Genomic sequencing of commensal microorganisms derived from multiple hosts provides opportunities to
model the dynamics of colonisation and within and between-host transmission in an ecosystem, without the
need for an infectious disease outbreak. Additionally, the pathogenic potential of generalist microorganisms
capable of infecting more than one host species gains importance when reservoirs act as source of infection
for spill-over to naive vulnerable species. A classic example is the emergence of squirrel pox virus (SQPV)
contributing to the decline of the native red squirrel (Sciurus vulgaris) in the United Kingdom. Although
transmission of the virus from the introduced grey squirrel (Sciurus carolinensis) has been demonstrated, the
origin of squirrel pox virus in grey squirrels still remains unclear (Rushton et al. 2006).
Island ecosystems provide the ideal environment for the investigation of bacterial transmission between host
communities. Range restriction imposed by geographical isolation can limit immigration and emigration of
hosts, with the notable exception of mobile species such as birds which often have extensive home ranges.
Conservation practices within New Zealand use mainland and offshore islands as sites for the protection of
multiple endangered species (Sherley et al. 2010), due to ease of biosecurity and predator control, thus
increasing diversity and density of potential hosts. Concentration of animals at unnaturally high densities on
isolated islands may create hubs for the intra- and inter-specific transmission of microorganisms between
wildlife communities. Host density and connectivity have been highlighted as influential factors determining
the transmission of Escherichia coli between ungulates in Kenya (VanderWaal et al. 2013b, 2014). The
translocation of animals between ecologically distinct locations may bring together previously allopatric
populations and species into sympatry, providing opportunities for pathogens to cross natural barriers
94
between host communities. If already vulnerable, any additional selective pressure imposed on extant naive
populations may lead to disease outbreaks. Although disease is not a common sole driver of population
decline (Heard et al. 2013; Smith et al. 2006), it may contribute to localised extinction. For example, the
spread of the chytrid fungus Batrachochytrium dendrobatidis along with environmental factors has resulted
in the decline and extinction of frog species around the world (Berger et al. 1998; Fisher et al. 2012).
Campylobacter spp. are enteric bacteria able to colonise a diverse range of hosts (Kwan et al. 2008),
including multiple avian species with variable prevalence reported depending upon species, location and
method used (Waldenstrom & Griekspoor 2014). Although Campylobacter spp. are often harmless
commensals in wildlife, they can be pathogens when hosts are exposed to isolates from another species. An
infection experiment demonstrated a marked decrease in the body mass of European robins (Erithacus
rubecula) following inoculation with a Campylobacter jejuni isolate derived from another wild bird species
(Waldenstrom et al. 2010). Previous studies have highlighted the exceptionally high prevalence of
Campylobacter spp. isolated from takahe (Porphyrio hochstetteri) (Chapter 3). Takahe are an endemic
endangered New Zealand flightless rail (BirdLife International 2013) whose populations have been heavily
manipulated through conservation actions. Preservation of the species relies on the frequent translocation of
individuals between remote sanctuaries (Wickes et al. 2009). This study describes the prevalence of
Campylobacter spp. in five vertebrate orders within an island ecosystem used for the conservation of takahe.
To explore bacterial transmission between takahe and potential reservoirs, comparative analyses of 52 of the
53 conserved ribosomal protein genes used in ribosomal multi-locus sequence typing (rMLST) schemes
(Jolley et al. 2012) were used for the discrimination and differentiation of Campylobacter spp. genomes
isolated from hosts on Maud Island, New Zealand. To our knowledge, this is the first study to use genomic
comparisons of rMLST genes to investigate the epidemiology of a commensal bacterium within abundant
and endangered communities inhabiting the same island ecosystem.
5.3. Methods
5.3.1. Study site
Maud Island, also known as Te Hoiere scientific reserve, is located in the Marlborough Sounds, South
Island, New Zealand. With a history of farming until it became a wildlife reserve in 1974, the island was
cleared of vegetation but through restorative efforts is now comprised of approximately 310 hectares of
predominantly regenerating coastal forest. The mainland is 900 metres away at the closest point. The island
95
is a predator-free conservation reserve and has a long history of translocations of endangered species,
including takahe, to the site.
5.3.2. Study populations
Individuals from five vertebrate orders on Maud Island (Table 5-1) were sampled and tested for
Campylobacter spp. via faecal or cloacal swabs in either a repeated or simple cross sectional design. Four
takahe were caught and sampled opportunistically in March 2013 during New Zealand Department of
Conservation (DOC) health checks or for management purposes. During the study period there were sporadic
incursions by weka (Gallirallus australis) onto the island. Samples were collected opportunistically from 59
weka, by DOC staff between January and May 2013 prior to relocation to the mainland. All takahe and weka
were uniquely identified by DOC administered leg rings.
Mist nets were erected for up to 3 days within takahe territories, in three month intervals between April 2012
and March 2013 in order to capture, sample and release fifty passerines per season. Nine birds were
recaptured in separate events and sampled. Recaptures were not included in prevalence estimates or
subsequent analysis. Birds were either marked with paper correction fluid (n = 26) or given a unique leg ring
identifier (n = 165). In April 2012, faecal swabs (n = 50) were collected from passerine holding bags. During
all subsequent capture periods, passerines were sampled by cloacal swabs (n = 150).
Due to seasonal fluctuations in presence on the island, 50 little blue penguins (Eudyptula minor) were
captured over three nights during the breeding season in October 2012. Individuals were caught by hand or
hand net, restrained for processing and subsequently released at the point of capture. All penguins were
microchipped subcutaneously in the midline dorsal scruff of the neck with a unique identifier in order to
prevent recapture. All penguins were sampled by cloacal swab prior to release.
In January 2013, 99 reptiles, comprising 80 common geckos (Hoplodactylus maculatus) and 19 brown
skinks (Oligosoma zelandicum), were caught by hand from under artificial and natural retreats. Reptiles were
temporarily marked using paper correction fluid on the dorsal surface to prevent resampling over the three
days sampling period and cloacal swabs were collected prior to release at point of capture.
Approximately 50 domestic sheep (Ovis aries) were present on the island. Sheep were observed within
paddocks and 100 fresh faecal samples were collected, over a period of two days in January 2013, with the
aim of sampling most individuals. Faecal samples were collected a minimum of one meter radius apart in
order to avoid cross contamination.
96
5.3.3. Sample collection
Cloacal or faecal swabs were collected from all species captured between April 2012 and May 2013 and
immediately placed into Aimes charcoal transport media (Copan, California, USA). Cloacal swabs were
collected and stored refrigerated at 4oC in Aimes charcoal transport media for up to 7 days.
5.3.4. Microbiological culture, molecular confirmation and speciation
Microbial culture for Campylobacter spp. and DNA extraction were conducted using previously described
methods (Chapter 3). Two colonies with typical Campylobacter spp. morphology were randomly selected
from positive modified Charcoal Cefoperazone Deoxycholate Agar (mCCDA) plates, followed by DNA
extraction using methods previously described (Chapter 3). The selection of 2 colonies was applied due to
study constraints; however this method limits the detection of multiple species potentially carried by an
individual, and thus the interpretation of subsequent analysis. Suspected Campylobacter spp. isolate DNA
was subject to confirmatory PCR using published protocols targeting: the mapA gene in Campylobacter
jejuni (Mullner et al. 2010; Stucki et al. 1995), the ceuE gene in Campylobacter coli (Denis et al. 2001;
Gonzalez et al. 1997), the putative C4-dicarboxylate trans-membrane transport gene in Campylobacter sp.
nova 1 (Chapter 3) and a specific region of the 16S rRNA gene to identify members of the genus
Campylobacter (Linton et al. 1997).
Apparent and true prevalence estimates with 95% confidence intervals (95% CI) of C. jejuni , C. sp. nova 1
and unidentified Campylobacter spp. were calculated in the epiR package (Stevenson 2014) in R software (R
Core Team 2013) using Blaker’s intervals (Blaker 2000) with estimates of sensitivity (se) and specificity (sp)
previously described: C. jejuni se = 62.9, sp = 95.7% and C. sp. nova 1 se = 69.1, sp = 97.8% (Chapter 3).
Sensitivity and specificity values for Campylobacter spp. culture and PCR were based on those for C. sp.
nova 1.
One isolate from each Campylobacter spp. positive individual was selected at random for whole genome
sequencing, unless the two isolates extracted from an individual were identified as different Campylobacter
spp. by PCR, in which case both were sequenced. One isolate was unable to be recovered from storage and
was not included in further genomic analysis. Genomic DNA preparation, sequencing, assembly, curation
and annotation were conducted as per protocols previously described in Chapter 4.
97
5.3.5. rMLST analysis
Nucleotide sequences for 52 of the 53 genes (the order Campylobacterales does not possess the rpmD gene)
encoding bacterial ribosomal protein subunits used for ribosomal multi locus sequence typing (rMLST)
(Jolley et al. 2012), were identified and extracted from assembled draft genomes using a C. jejuni reference
set of rMLST genes as a database for a BLAST search. Using custom Perl scripts, unique alleles for each
gene were called based on their nucleotide sequence. Each individual gene was aligned using Muscle (Edgar
2004) and the aligned genes were concatenated to make a single alignment per genome. A NeighborNet tree
of the 52 concatenated rMLST gene nucleotide sequences derived from the 19 Campylobacter spp. genomes
was visualised in SplitsTree4 (Huson & Bryant 2006). Pairwise comparison of genomic distance between
Campylobacter spp. was conducted within SplitsTree4 (Huson & Bryant 2006) using uncorrected P
measures. This distance is obtained by dividing the number of single nucleotide differences by the total
number of nucleotides compared between two gene sequences. The lower triangle distance matrix was
visualised using the pheatmap package (Kolde 2013) in R software (R Core Team 2013).
5.3.6. In silico PCR of the C4-dicarboxylate trans-membrane transport gene
In silico C. sp. nova 1 PCR was conducted on all sequenced isolates in order to test for the presence of the
C4-dicarboxylate trans-membrane transport gene and nucleotide polymorphisms within the target sequence.
This was conducted because some isolates were identified as closely related to C. sp. nova 1 by rMLST but
tested negative by in vitro PCR. Forward and reverse primer sequences for the C4-dicarboxylate trans-
membrane transport gene in Campylobacter sp. nova 1 (Chapter 3) were used to find the PCR target region
of one isolate (Appendix 9.5-1). The 106 base pair target region was extracted and used as a reference to
BLAST against all other isolates in order to determine the presence of the gene within a given genome and
any nucleotide polymorphisms which may be present. Sequences were aligned and visualised within
Geneious v7.17 (Drummond et al. 2013).
5.4. Results
5.4.1. Prevalence of Campylobacter spp. in vertebrate communities
Five vertebrate orders, including 294 birds, 99 reptiles and approximately 50 sheep, inhabiting Maud Island
were tested for Campylobacter spp. between April 2012 and March 2013. Sampling details and
Campylobacter spp. apparent (AP) and true (TP) prevalence estimates are presented in Table 2-1. All four
takahe tested positive for Campylobacter with two species isolated; C. jejuni (TP, = 35.3% [95% CI 0-98
100%]) and C. sp. nova 1 (TP = 100% [95% CI 68-100%]). Campylobacter spp. were isolated from eleven
weka tested during the island incursion. Weka prevalence estimates were as follows; C. sp. nova 1 (TP =
17.0% [95% CI 6-33%]) and unidentified Campylobacter spp. belonging to the Campylobacteraceae family
(TP = 4.3% [95% CI 0-17%]). Season was not associated with Campylobacter spp. carriage in the passerines
tested. Two species of Campylobacter were isolated from passerines, including C. jejuni isolated from one
bellbird (Anthornis melanura) (TP = 0% [95% CI 0-5%]) and two silvereyes (Zosterops lateralis) (TP = 0%
[95% CI 0-10%]). We identified an unknown species of Campylobacter from a Eurasian blackbird (Turdus
merula) (TP = 13.3% [95% CI 0-63%]). Campylobacter spp. was not isolated from any of the reptiles, little
blue penguins or sheep tested during the study period.
5.4.2. Comparative genomics of Campylobacter spp.
Nineteen Campylobacter spp. isolates, encompassing all but one Campylobacter spp. positive individual due
to the inability to recover a weka derived C. sp. nova 1 isolate from storage, were subject to whole genome
sequencing. Four distinct clusters were identified based on pairwise comparison of rMLST nucleotide
sequences. In addition, potential recombination within and between Campylobacter spp. was evident from
the net-like structure in some parts of the NeighborNet phylogeny (Figure 5-1A), indicative of uncertainty in
the phylogenetic relationships between isolates. The largest genetic distances were observed between
different Campylobacter spp., with some evidence of host associated grouping of related sequence types
within a Campylobacter spp. (Figure 5-1B). Takahe C. sp. nova 1 sequence types were highly related to each
other and more distantly to three of the C. sp. nova 1 isolated from weka within cluster one (Figure 5-1A).
The remaining four weka C. sp. nova 1 isolates were positioned within cluster two next to unidentified
Campylobacter spp. isolates also isolated from weka. The distance between the two clusters of C. sp. nova 1
was larger (pairwise uncorrected P measure value, PV = 0.1) than that observed between the unidentified
Campylobacter spp. and C. sp. nova 1 within cluster 2 (PV = 0.01-0.03) (Figure 5-1B).
C. jejuni isolated from passerines (bellbird and silvereyes) were more closely related to each other (PV =
0.001 - 0.01) than to the one takahe associated C. jejuni (PV = 0.02) (Figure 5-1B). All C. jejuni isolates
were genetically separated from C. sp. nova 1, as indicated by the long branch lengths in the NeighborNet
tree. The unknown Campylobacter spp. isolated from a Eurasian blackbird was genomically distinct with no
shared allele sequences (Figure 5-1, Table 5-2) and a large genetic distance (PV = 0.14 - 0.15) (Figure 5-1B)
from all other Campylobacter spp. isolated from vertebrates on Maud Island in this study.
99
Species Scientific name No. tested (No. positive) Campylobacter spp. Campylobacter spp. prevalence (95% CI) Apparent (AP) True (TP) Passerines Bellbird
Anthornis melanura
69 (1)
C. jejuni
1.5% (0-8%)
0 % (0-5%)
Chaffinch Fringilla coelebs 3 - - - Dunnock Prunella modularis 3 - - - Eurasian blackbird Turdus merula 11 (1) Campylobacter spp. 11.1% (1-44%) 13.3% (0-63%) European goldfinch Carduelis carduelis 1 - - - Grey warbler Gerygone igata 2 - - - House sparrow Passer domesticus 22 - - - New Zealand fantail Rhipidura fuliginosa 8 - - - Silvereye Zosterops lateralis 66 (2) C. jejuni 3.0% (1-10%) 0% (0-10%) Song thrush Turdus philomelos 2 - - - Tui Prosthemadera novaeseelandiae 4 - - - Reptiles Brown skink
Oligosoma zelandicum
19
-
-
-
Common gecko Hoplodactylus maculatus 80 - - - Rails Takahe
Porphyrio hochstetteri
4 (4)
C. jejuni C. sp. nova 1
25.0% (1-75%) 100% (47-100%)
35.3% (0-100%) 100% (67-100%)
Weka Gallirallus australis 59 (11*) Campylobacter spp. C. sp. nova 1
5.1% (1-14%) 13.6% (6-24%)
4.3% (0-17%) 17.0 (6-33%)
Seabird Little blue penguin
Eudyptula minor
50
-
-
-
Ruminant Sheep
Ovis aries
100 faecals
-
-
-
*One C. sp. nova 1 isolate was unable to be recovered for sequencing One takahe tested positive for both C. jejuni and C. sp. nova 1.
Table 5-1 List of species captured and sampled for Campylobacter spp. on Maud Island, New Zealand, including sampling effort, apparent (AP) and true prevalence (TP) of Campylobacter spp. identified by in vitro PCR.
100
Figure 5-1 NeighborNet tree (A) and pairwise genomic distance matrix (B) based on 52 gene rMLST nucleotide sequences using uncorrected P values (PV) of Campylobacter spp. isolated from host species on Maud Island.
101
Table 5-2 Allelic profiles of 52 rMLST genes present within Campylobacter spp. isolated from host species on Maud Island, New Zealand.
102
In silico PCR for the putative Campylobacter spp., C. sp. nova 1 was conducted on all Campylobacter spp.
positive isolates and confirmed the presence of the target gene within all in vitro PCR positive C. sp. nova 1
(Appendix 9.5-1). One hundred percent sequence homology was observed between 7 out of 11 identified
genes. Three isolates originating from weka contained 7 single nucleotide polymorphisms (SNPs) and one
weka isolate possessed 4 SNPs within the target region (Appendix 9.5-1). There was no match to the PCR
target gene within isolates identified as C. jejuni or unidentified Campylobacter spp.
Evidence of recombination and sharing of identical alleles was identified on comparison of allelic profiles
within and between hosts. This was also observed among Campylobacter spp. (Table 5-2). Identical alleles
were predominantly nested within a Campylobacter spp. Alleles were often conserved within a host with
occasional sharing of identical sequences between hosts. Interestingly, five genes (rplN, rpmC, rpsH, rpsM
and rpsS) showed evidence of homologous sequences within both C. sp. nova 1 and C. jejuni isolated from
takahe, and 8 alleles were shared between C. sp. nova 1 and unidentified Campylobacter spp., indicating
potential recombination between Campylobacter spp. Some genes were relatively conserved across both
bacterial species and host. For example, only three sequence combinations were defined for the rpmJ gene
across the 19 isolate genomes.
5.1. Discussion
Management of existing and emerging wildlife disease threats in endangered or threatened populations
requires an understanding of host-pathogen relationships and transmission within natural ecosystems.
Previous studies investigating the bacterial dynamics of a common enteric bacterium, Campylobacter spp.
within fragmented takahe populations discovered a complex relationship between host location and bacterial
carriage (Chapter 3 ; Chapter 4). This raised the question of whether location attributes including direct and
indirect interactions with sympatric reservoir hosts contributed to the prevalence and bacterial sequence
types harboured within takahe. Results from this small-scale cross-sectional prevalence survey provide
evidence that avian species, in particular members of the Rallidae family, are the predominant host of
Campylobacter spp. within the study island ecosystem. We observed evidence of allele sharing between
different Campylobacter spp. isolated from the same and taxonomically distinct hosts suggesting a recent
shared evolutionary history of some genes, most likely the result of genetic recombination between these
bacteria. Transmission of a bacterium between sympatric hosts may be inferred from the bacterial sequence
type they carry (Chiyo et al. 2014; VanderWaal et al. 2013b, 2014). Clonal sequence types were not detected
in this study, thus transmission between hosts through direct or indirect means cannot be inferred. However,
103
conservation of many rMLST genes was found to be within a host species and those that were shared tended
to be between species of the same family order, suggesting cross-species transmission would most likely
occur between closely related hosts.
Campylobacter spp. are a multi-host enteric organism which has been isolated from a range of species
including companion animals (Mughini Gras et al. 2012), wildlife (Colles et al. 2011; French et al. 2009;
Keller & Shriver 2013; Mohan et al. 2013; Sippy et al. 2012) and farm animals (Colles et al. 2011; Kwan et
al. 2008). Although Campylobacter spp. isolated from takahe is thought to be of a commensal nature in
takahe with a high prevalence reported in this and previous studies (Chapter 3 ; Chapter 4), exposure to
different host associated genotypes may result in reduced fitness and pathogenic effects in the new host
(Waldenstrom et al. 2010). We report a low prevalence of Campylobacter spp. in sympatric communities
sharing an island with takahe. However, prevalence estimates do not necessarily correlate with the likelihood
of transmission of a microorganism to new hosts. The ecological concept of “super-spreaders” and “super-
shedders” (Lloyd-Smith et al. 2005) based on the 20:80 rule (Woolhouse et al. 1997; Woolhouse et al. 2005),
implies that a small proportion of the population infected with a microorganism may be responsible for the
dissemination of pathogens to new hosts. Therefore, although the prevalence of Campylobacter spp.
observed in species co-habiting an environment with takahe was low in comparison to that observed in
takahe, it is theoretically possible that the few infected individuals may significantly contribute to the
transmission of Campylobacter spp. between hosts.
Campylobacter spp. was only detected in terrestrial birds, with a higher prevalence in the flightless takahe
and weka. Environmental and behavioural variables may influence transmission between cohabiting
vertebrate communities. Faecal contamination of the environment provides opportunities for mixing and
chance transmission of enteric bacteria between taxonomically distinct hosts sharing an environment.
However, transmission through faecal oral pathways requires a period of survival in the environment (Bull et
al. 2006) and given the relative fragility of Campylobacter spp. and sub-optimal conditions outside the host
(Murphy et al. 2006), it seems likely that mixed species co-infection may be an important means of
horizontal gene transfer between Campylobacter spp. Similarities in host behaviour, structuring and foraging
between closely related hosts may explain the carriage and genetic similarities between C. sp. nova 1 isolated
from takahe and weka. This concept is supported by other studies where host exposure to Campylobacter
spp. has been attributed to ecological guild with a higher incidence reported in ground foraging and
scavenging birds (Waldenstrom et al. 2002). In addition, genomic comparisons between C. sp. nova 1
isolated from weka with those from takahe in different locations (unpublished data) identified identical 104
rMLST profiles between an isolate from a takahe in the Murchison Mountains, New Zealand and a C. sp.
nova 1 from a weka in this study.
Bacteria that colonise multiple hosts, with variable pathogenicity in different hosts, are likely to exhibit
heterogeneity in carriage between host species due to differences in host immunity, population sizes, social
behaviour, spatial distribution and habitat use within an ecosystem (Chiyo et al. 2014; Dobson 2004; Grange
et al. 2014). Campylobacter spp. were not isolated from sheep on Maud Island during this study, although
sheep are known carriers of Campylobacter spp. (Mughini Gras et al. 2012; Mullner et al. 2009).
Additionally, a recent study found close associations between proximity to agricultural environments and
increased likelihood of carriage of Campylobacter coli carriage in takahe (Chapter 3). Therefore, co-
habitation of livestock with species of conservation concern should only proceed with caution and strict
biosecurity including a quarantine period and health testing. Foraging behaviour of marine birds is
substantially different from terrestrial species, with penguins spending substantial amounts of time at sea in
pursuit of food only returning to shore at night to rest within nests or burrows, thus reducing opportunities
for faecal-oral transmission and colonisation of Campylobacter spp. C. jejuni has been isolated from
macaroni penguins (Eudyptes chrysolophus) in the Antarctic, however this is a rare report and the incidence
was thought to be associated with human effluent contaminating the penguin habitat (Griekspoor et al. 2009).
Most reports of Campylobacter spp. isolation in reptiles are from captive animals. For example,
Campylobacter fetus and Campylobacter hyointestinalis were isolated from 44.8% of captive reptiles tested
in a study in the Netherlands (Gilbert et al. 2014). However, carriage of Campylobacter spp. in wild reptilian
populations is rare, as supported by this study.
The aim of this study was to investigate potential reservoirs of Campylobacter spp. which may be a source of
infection in takahe. Interestingly, with such a high incidence unmatched in other animal communities tested,
the endangered takahe may themselves be a sustainable source of infection to other vulnerable species.
Takahe on Maud Island form part of a complex highly connected network of takahe populations residing
within conservation reserves (Grange et al. 2014). Linked through regular translocations, localised and long
range transmission through faecal contamination of the environment may be possible. In New Zealand,
islands are commonly used as refuges for multiple translocated species of conservation concern (Sherley et
al. 2010). Transmission of the apparently rail associated C. sp. nova 1 (Chapter 3 ; Chapter 4 ; French et al.
2014) to naive hosts or humans interacting with takahe may be detrimental, although the pathogenicity of
this species in other hosts is unknown.
105
Detection and identification of Campylobacter spp. in biological communities contributes only a small
amount to the understanding of the complex interactions involved in the epidemiological triangle of
interactions between the host, microbe and the environment. This study was limited by the relatively small
populations and sample sizes, and the practical need for opportunistic cross sectional sampling often
encountered in investigations of wildlife disease ecology. An inherent risk of ecological studies is there may
be a range of unmeasured biotic and abiotic factors influencing the individual through to population and
community scales (Tompkins et al. 2011), which may explain the host-microbe relationships observed in this
study. For example, survival and establishment of Campylobacter spp. within a host may be dependent upon
interactions with other members of the gut microbiome, which may in turn be confounded by factors such as
host age and ecological interactions (Costello et al. 2012). Additionally, it is likely that this study did not
encompass the full diversity of Campylobacter spp. present within the study species due to small sample
sizes, low prevalence of carriage and the test characteristics of the methods used. Previous comparisons of
culture versus faecal DNA detection revealed imperfect sensitivity and specificity of the culture method used
for Campylobacter spp. detection (Chapter 3). Epidemiological investigations using microbial genomics
require bacterial culture of an isolate prior to sequencing and analysis. Therefore, studies of this kind need to
take into account the biases resulting from the selective nature of culture which will have downstream effects
on the interpretation of results.
Genomic analysis using rMLST profiles seemed to contradict the species definition of C. sp. nova 1
previously described in studies of the bacterium in takahe (Chapter 3 ; Chapter 4), due to the relatively large
distances between the isolates within cluster 1 and 2, as well as the close relationship between C. sp. nova 1
and unidentified Campylobacter spp. within cluster 2 (Figure 5-1). Although in silico PCR confirmed the
presence of the C. sp. nova 1 PCR target within all C. sp. nova 1 isolates identified by in vitro PCR (Chapter
3), rMLST analysis revealed distinct genomic differences between C. sp. nova 1 isolates (Figure 5-1). A
group of C. sp. nova 1 isolates were closely related to unidentified Campylobacter spp. isolates which do not
possess the PCR target gene. The absence of this gene within the unidentified Campylobacter spp. may have
been a result of gene deletion or absence of insertion. Alternatively, given the genetic distance between
cluster one and two, C. sp. nova 1 within cluster two may be a different species which have acquired or
retained the PCR target gene from recombination with an ancestor of the isolates in cluster 1. Either way,
this example highlights the restrictive nature of using conventional PCR methods for bacterial species
detection and identification, as well as the need to use a large gene pool when comparing bacterial isolates
for epidemiological investigations of bacterial transmission or source attribution.
106
Management of wildlife infectious disease requires the identification, characterisation and quantification of
microorganisms in order to understand mechanisms which drive transmission and persistence. This small-
scale study demonstrates the use of population genetic tools to explore complex host associated bacterial
carriage. High resolution strain typing and determination of host association through genomic sequencing,
provides more robust evidence and confidence previously unattainable through conventional molecular
methods. Modelling of commensal bacteria in this study provides insight into potential epidemiological
interactions between hosts and pathogens, including the roles that ecologically similar and distinct hosts may
play in epidemics and the transmission of more pathogenic generalist microorganisms, such as Salmonella
spp.
5.2. Acknowledgements
This study was funded by the Allan Wilson Centre. We would like to thank Ngati Kuia and the Department
of Conservation for their support and encouragement in this study. We would like to acknowledge C. and L.
Birmingham, T. Burns, M. Jensen, S. Michael, D. Sijbranda, P. Njiman for field assistance and A. Reynolds
for laboratory support.
5.3. Animal ethics and permits
Samples were collected under a Massey University animal ethics permit MUAEC Protocol 11/95. Capture,
handling and sample collection were in accordance to New Zealand Department of Conservation permits
NM-33051-FAU and NM-35424-FAU. Passerines were caught and banded under New Zealand national
banding scheme permit number 2012/007.
5.4. Supplementary information
The sequence alignment for the in silico Campylobacter sp. nova 1 target region (Appendix 9.5-1) is
available in the appendix.
107
108
CHAPTER 6
LOCATION SPECIFIC PREVALENCE OF SALMONELLA SPP. IN
ENDANGERED TAKAHE (PORHPYRIO HOCHSTETTERI)
109
110
6. Location specific prevalence of Salmonella spp. in endangered takahe
(Porphyrio hochstetteri)
6.1. Abstract
Generalist opportunistic bacteria such as Salmonella spp. are able to cross transmission boundaries between
host species. The process of translocation of endangered species to new locations may create novel routes of
transmission of potentially pathogenic Salmonella spp. between previously isolated hosts and reservoirs,
with the potential for negative impacts on vulnerable populations. We investigated the prevalence of
Salmonella spp. in endangered takahe (Porphyrio hochstetteri) from multiple geographically isolated
locations in New Zealand. A low Salmonella spp. prevalence, 1-5% (95% CI 0-16%), was detected, with
three serotypes isolated; S. enterica subsp. enterica serotype Mississippi, S. enterica subsp. enterica serotype
Saintpaul and S. enterica subsp. houtenae serotype 40g.t. Takahe from a single island location were
significantly (p=0.002) more likely to carry Salmonella spp. at the time of sampling than any other location.
The serotypes isolated from takahe on this island have been associated with reptiles in other studies, and S.
enterica subsp. enterica serotype Saintpaul is an important zoonotic serotype in New Zealand. The
geographic clustering of positive cases suggests the presence of environmental reservoirs and transmission
routes on the island which are unavailable elsewhere. The physiological impact of Salmonella spp. infection
on takahe is unknown; however translocation of Salmonella spp. positive individuals to new sites presents a
risk of disease expression within the host as well as transmission to other species.
111
6.2. Introduction
Animals are host to many different organisms which may be commensals, symbiotic, parasitic or pathogenic
in nature. Salmonella spp. are opportunistic pathogens, able to infect a broad range of hosts, raising concerns
about zoonotic transmission through foodborne (EFSA 2013) and direct routes (Cummings et al. 2012) .
Salmonella spp. infection can have negative impacts on wildlife populations (Hall & Saito 2008; Lawson et
al. 2010). Geographical isolation of endangered species, habitat fragmentation and movements of animals to
new environments presents opportunities for exposure to novel and potentially pathogenic microorganisms.
Therefore, increased understanding of the factors that influence the biological package of parasites and
pathogens within hosts and their communities is required to understand the implications of conservation
actions such as translocation on the spread of pathogens (Bengis et al. 2002).
Fauna which are translocated to a new environment may pose a significant disease risk to the incumbent wild
populations. Animals can act as carriers and / or vectors of exotic pathogens into naive wild populations
which have no effective immunity to these pathogens at the release site (Anderson & May 1986; Woodford
1993). Awareness of this issue has prompted protocols to minimise disease transmission risk associated with
wildlife translocations and reintroductions (Cunningham 1996; Viggers et al. 1993; Woodford & Rossiter
1993). Due to such concerns, health screening is now required prior to translocations of the endemic
endangered New Zealand flightless takahe (Porphyrio hochstetteri) (BirdLife International 2013; McInnes et
al. 2004). Included in the health test profile is culture for Salmonella spp. from cloacal samples. Although
useful, interpretation of results remains difficult due to imperfect tests, intermittent shedding of faecal
Salmonella (Ivanek et al. 2012; Van Immerseel et al. 2004), and the timing of testing. Encounters with new
pathogens may occur at several stages of the translocation process (Cunningham 1996). The prevalence of
Salmonella spp. infection in wild birds is reportedly low, often associated with asymptomatic carriage,
however outbreaks of salmonellosis have caused morbidity and mortality in passerines throughout the world,
including the USA, United Kingdom and New Zealand (Alley et al. 2002; Hall & Saito 2008; Lawson et al.
2010; Pennycott et al. 2010). Unfortunately there is a lack of baseline information on Salmonella spp.
prevalence in wild populations of takahe, with anecdotal reports of Salmonella enterica subsp. enterica
serotype Brandenburg (Orr 1997), Salmonella enterica subsp. enterica serotype Typhimurium (McLelland et
al. 2011) and Salmonella Mana (ESR 2010) isolated from a few individuals. We investigated the prevalence
and geographic distribution of Salmonella spp. carriage in takahe from multiple locations in New Zealand.
112
6.3. Methods
One hundred and forty seven individual takahe, and 15 repeat samples of individuals which had been
translocated to a new location in the time frame of this study were tested opportunistically during pre-
translocation disease screening or annual health checks. Takahe were tested for cloacal and / or faecal
carriage of Salmonella spp. from nine locations (Table 6-1) within New Zealand between November 2011
and April 2013. This sample size represents approximately 66% of the total population of this critically
endangered species and covers most of its remaining geographic range. For reasons of confidentiality, the
location of the private island is undisclosed. All swabs were transported and stored in Aimes Charcoal
Transport media (Copan, California, USA) at 4oC for one to seven days prior to culture.
Culture for Salmonella spp. was in aerobic conditions at 37oC unless otherwise stated. Each swab was
suspended in 2ml Phosphate buffered saline (PBS, pH 7.3). A 500µl aliquot of PBS solution was transferred
into 20ml buffered peptone water (BPW) (BioRad, Auckland, New Zealand) and incubated for 24 hours for
pre-enrichment. Subsequently, 100µl of PBS was inoculated into two secondary selective enrichment broths,
10ml Rappaport-Vassiliadis Salmonella (RVS) at 42oC and 10ml tetrathionate (TET) broth (Fort Richards,
Auckland, NZ) enriched with 100µl iodine-iodide solution (0.25g KI, 0.3g I/ml H2O). Both RVS and TET
were incubated for 24 +/- 2 hours. Incubated RVS and TET solutions were subcultured and streaked onto
Xylose Lysine Deoxycholate (XLD) and Brilliant Green Modified (BGM) agar (Fort Richards, Auckland,
NZ), and were incubated for 18 to 24 hours. The agar plates were examined and two to four suspect
Salmonella colonies of differing morphology were subcultured onto MacConkey agar (Fort Richards,
Auckland, NZ) and incubated for 18 to 24 hours. Cultures presenting as grey on MacConkey plates were
inoculated into Triple Sugar Iron (TSI) and Lysine Iron Agar (LIA) slopes (Fort Richards, Auckland, NZ)
and incubated for 18 to 20 hours. Isolates with positive reactions on the slopes were further tested for oxidase
reactivity and reaction to polyO and polyH antisera (Oxoid, Auckland, NZ). Oxidase negative and polyO and
or polyH antisera positive isolates were cultured on to Columbia Horse Blood Agar (Fort Richards,
Auckland, NZ) for 24 hours prior to identification using a RapID One test (Oxoid, Auckland, New Zealand)
conducted according to manufacturer’s instructions. Isolates identified as Salmonella spp. were cultured on
Dorset Egg agar slopes (Fort Richards, Auckland, NZ) for 24hrs and sent chilled to the Environmental
Science and Research (ESR-NCBID) laboratories at Wallaceville, New Zealand for serotyping.
Prevalence calculations were conducted using the EpiTools package (Aragon 2012) within R software (R
Core Team 2013). Estimates of the true prevalence, and corresponding 95% confidence intervals, of
113
Salmonella spp. in the takahe population were made using the test prevalence and assumptions that the
culture technique was 50% sensitive and 98% specific based on previous studies (Bager & Petersen 1991).
Prevalence estimates were calculated for different sample types, or combinations of sample types (positive
for either swab). Influence of location on Salmonella spp. presence and absence were performed using the
Fisher’s exact test in R software (R Core Team 2013).
6.4. Results
Three serotypes of Salmonella spp. were isolated from takahe (Table 6-1). Five isolates of Salmonella
enterica subsp. enterica serotype Mississippi and one of Salmonella enterica subsp. houtenae serotype 40g.t
(1) were isolated from takahe located on a private island. A takahe which had previously tested positive for
S. Mississippi on the private island was later found to be positive for Salmonella enterica subsp. enterica
serotype Saintpaul after translocation to the Burwood Bush breeding centre.
Location n (n positive) Salmonella serotype Burwood bush breeding centre 59 (1) Salmonella enterica subsp. enterica serotype
Saintpaul (1)* Mana Island 6 - Maud Island 4 - Maungatautari reserve 8 - Murchison Mountains 46 - Private Island 19 (6) Salmonella enterica subsp. enterica serotype
Mississippi (5) Salmonella enterica subsp. houtenae serotype 40 g.t. (1)
Te Anau wildlife reserve 4 - Tiritiri Matangi Island 15 - Willowbank reserve 1 -
*One takahe (Porphyrio hochstetteri) individual was translocated and previously tested positive for Salmonella Mississippi on the private island Table 6-1 Number (N) of takahe (Porphyrio hochstetteri) tested by location with corresponding Salmonella spp. serotypes isolated from takahe.
Six out of the seven Salmonella spp. positive individuals were identified from cloacal swab culture. Of these,
four birds had corresponding faecal swabs, two of which were in agreement with cloacal swab results, and
two of which were negative for Salmonella spp. One S. Mississippi isolate was detected in a faecal swab but
its corresponding cloacal swab was negative.
Overall true prevalence estimates for Salmonella spp. in takahe are relatively low, ranging between 1 and 5%
(95% CI 0-16%) (Table 6-2). Significant differences in Salmonella spp. prevalence were found according to
location (p-value 0.002), with Salmonella spp. positive individuals only detected in two locations, with true
114
prevalence estimates of 62% (private island, 95% CI 26-100%), and 0% (Burwood Bush breeding centre
95% CI 0-1%).
Sample type n (n positive) Salmonella true prevalence (95% CI) Cloacal swab 147 (6) 4% (0-14%) Faecal swab 130 (3) 1% (0-9%) Faecal or cloacal swab 162 (7) 5% (0-14%)
Table 6-2 Estimates of true prevalence of Salmonella spp. in takahe (Porphyrio hochstetteri) using three different sampling methods.
6.5. Discussion
Salmonella spp. are often regarded as opportunistic pathogens, which may be carried asymptomatically by
healthy individuals. However, recrudescence and increased shedding may occur during times of stress, such
as that induced by translocation (Verbrugghe et al. 2011). Monitoring two thirds of the takahe population for
intestinal carriage of Salmonella spp. over a one and half year time period revealed a low prevalence of
Salmonella spp., with a geographic clustering of asymptomatic Salmonella spp. carriage in takahe on a
private island.
Takahe populations are highly fragmented, with an extant population of approximately 230 individuals
(Wickes et al. 2009). Population management and mitigation against inbreeding effects (Jamieson et al.
2006) has been heavily reliant on translocation of the birds to environments that are substantially different
from the founding ecosystem. The isolation of Salmonella spp. predominantly from one location, and the
serotypes identified in this study, suggests there may be an environmental reservoir and transmission route
on the island, not present elsewhere, from which takahe are being exposed to as an incidental host. However
the nature of the environmental reservoirs and the epidemiology of infection were not examined in this study.
Herpetofauna are often implicated as sources of Salmonella spp., indeed, S. Mississippi has been isolated
from reptiles in Australia and New Zealand (Ball 1991; Middleton et al. 2014), with some human cases
reported in Tasmania correlated to contact with wildlife (Ashbolt & Kirk 2006). Additionally, S. houtenae
40g.t (isolated in this study) was recently detected in lizards sampled on New Zealand offshore islands
(Baling et al. 2013). Takahe are omnivorous; although grasses comprise the majority of their diet,
opportunistic consumption of reptiles and insects has been observed, thus presenting a potential direct
transmission route for Salmonella spp. infection. Additionally, artificial feeding and water sources provide
congregation sites which increase contact rates between multiple hosts and their faeces. For example,
aggregations of European starlings (Sturnus vulgaris) have been implicated as sources of Salmonella spp. in
115
concentrated animal feeding stations (Carlson et al. 2011), and the role of introduced species such as house
sparrows as reservoirs of Salmonella spp. is well established (Alley et al. 2002).
Against veterinary advice, a takahe which tested positive for Salmonella spp. at the time of translocation was
moved from the private island to the breeding centre. One week later when retested in breeding centre, the
takahe was positive for S. Saintpaul, an unrelated serotype (Table 6-1). Given the short time period between
sampling and limitations of Salmonella culture, it may be possible the bird was carrying multiple serotypes
of Salmonella spp. at both time points. Salmonella Saintpaul is a common serotype and has been isolated
from a range of taxa, including wildlife (Iveson et al. 2009; Middleton et al. 2014; Parsons et al. 2010), and
humans (Munnoch et al. 2009). Therefore, an alternative explanation may be that the takahe was exposed to
S. Saintpaul during translocation or at the new site and the two serotypes isolated represent separate infection
events. It is worth noting the translocation of a Salmonella spp. positive individual into a breeding centre, a
site identified as a dispersal hub and important in terms of connectivity to the takahe population network
(Grange et al. 2014), poses a substantial risk of spread of Salmonella spp. to takahe within the location, and
forward transmission to other populations. With known relative stability in faeces, soil and water for up to
and over a year, Salmonella spp. may remain viable in the environment for long durations (Winfield &
Groisman 2003). Therefore, transmission of Salmonella spp. may occur from environmental sources long
after an individual has stopped shedding.
Asymptomatic carriage of Salmonella spp. in apparently healthy takahe in this study suggests that, in these
cases, there was a non-pathogenic host-microbe association. However, the opportunistic pathogenicity of
Salmonella spp. and its association with reproductive and enteric disease in many species suggests that
caution should be used in extrapolating these cases to all serotypes of Salmonella in takahe. The process of
capture and translocation can have health impacts on the host. For example, the process of translocation
could result in significant physiological stress of individuals making them vulnerable to disease outbreaks, as
occurred with Erysipelothrix rhusiopathiae outbreaks in translocated kakapo (Strigops habroptilus) (Gartrell
et al. 2005). Additionally, reserves are also home to many other threatened species, many of which are also
subject to translocation regimes. Therefore cross-species transmission is a substantial concern where
pathogenicity and impact of infection is unknown. Further work is required to understand the ecology and
epidemiology of Salmonella spp. in the island ecosystem to determine source attribution of Salmonella spp.
isolated from takahe, which may lead to mitigation and prevention of transmission of Salmonella spp. to
takahe.
116
6.6. Acknowledgements
Funding was provided by the Allan Wilson Centre. Samples were collected under a Massey University
animal ethics permit MUAEC Protocol 11/95. We would like to thank P. Marsh, G. Greaves, A. Wilson, B.
Jackson and the Friends of Tiritiri Matangi, the Department of Conservation and the Maori community for
their support.
117
118
CHAPTER 7
GENERAL DISCUSSION
119
120
7. General Discussion
7.1. Microbial dynamics in translocated takahe (Porphyrio hochstetteri)
This study was the first of its kind to explore microbial dynamics in a large proportion of a well-described
but fragmented population of a critically endangered bird, the takahe (Porphyrio hochstetteri). Sub-
populations are connected via a network of movements resulting from historical translocation events. Here I
discuss our ability to capture and quantify the epidemiology of infectious organisms resulting from the
translocation and isolation of wildlife populations using a case study of microbial dynamics in translocated
takahe (Porphyrio hochstetteri). The effects of translocation and isolation of takahe were investigated using
an array of different techniques and revealed new insights into host-microbe relationships resulting from
translocations.
I used a novel approach to explore the history of movements between remote populations of takahe. This
study revealed a complex network of sub-populations which were likely to vary in their propensity to spread
and maintain pathogens (Chapter 2 ; Grange et al. 2014). After describing the conservation network with
descriptive social network principles, I carried out empirical investigations using a commensal enteric
bacterium, Campylobacter spp., to explore meta-population microbial dynamics by sampling takahe in
multiple locations. Traditional multivariate and innovative genomic investigations identified that population
isolation and conservation management of the takahe hosts has influenced the carriage and population
structure of Campylobacter spp. genotypes (Chapter 3 ; Chapter 4).
My study suggests the intensive conservation management of takahe that resulted in a range expansion
following a significant bottleneck (Ballance 2001) has had unforeseen impacts on host-microbe relationships
and transmission dynamics (Torchin & Mitchell 2004). The management of takahe in different
environmental settings has influenced the carriage of Campylobacter jejuni and Campylobacter coli (Chapter
3). A newly discovered rail-associated Campylobacter sp. nova 1 was prevalent in all populations (Chapter
3). However, more discriminatory whole genome analysis of isolates detected a significant biogeographic
variation in C. sp. nova 1 genotypes (Chapter 4). Possible explanations for the observed pattern include
spatial expansion and isolation of hosts resulting in reduced gene flow of Campylobacter spp. and allopatric
speciation, the presence of heterogeneous environmental attributes or cross-species transmission of
Campylobacter spp. from reservoir hosts in the same location. An assessment of other vertebrate reservoirs
in the island ecosystem indicated cross-species transmission of Campylobacter spp. may be a factor
121
contributing to the maintenance and phylogeographical distribution of Campylobacter spp. in takahe. Indeed,
if spill-over of Campylobacter spp. were to occur, it would most likely transmit between takahe and other
closely related hosts, such as weka (Gallirallus australis) (Chapter 5). Geographical variations in the
genotypes of Campylobacter spp. carried by the host, suggests isolation and environmental variables are
influencing the ongoing evolution of Campylobacter spp. Isolation can therefore allow continued evolution
of the Campylobacter spp. carried by takahe, which may allow a change in state from their current benign
nature, to more pathogenic strains that might affect their avian hosts (Waldenstrom et al. 2010) or their
human caretakers.
Molecular detection and genomic analysis of Campylobacter sp. nova 1 has revealed new insights into the
taxonomy of the genus Campylobacter. This study discovered previously undetected diversity within the
putative species and identified a relatively close relationship between C. sp. nova 1 and the two human
pathogenic species: C. coli and C. jejuni (Chapter 4). Evidence of recombination between different
genotypes within the species (Chapter 4) and between known and unidentified Campylobacter species
(Chapter 5), suggests a complex ancestry and calls into question our definitions of Campylobacter species.
Geospatial clustering of C. sp. nova 1 genotypes, combined with the relatively recent translocation events,
provides evidence to support a theory that the species is able to adapt and evolve within a relatively short
time frame in an isolated host.
Given evidence of host association of C. sp. nova 1 genotypes (Chapter 5 ; French et al. 2014), it is likely
that ongoing relocation of takahe provides a route for co-infections and recombination of alleles between
genotypes explaining the present day genomic structure of isolates (Chapter 4). Trans-boundary transmission
and carriage of C. sp. nova 1 is supported by the isolation of a diverse range of genotypes from the breeding
population (Chapter 4). Further study is required to explain the strong association between population
connectivity, location and C. sp. nova 1 genotypes. However, I hypothesise historic and current management
practices are having unforeseen influences on enteric microbes, the consequences of which are unknown but
may be detrimental to the health of translocated populations of takahe. This includes exposure to strains of
micro-organisms from food producing animals that may result in increased pathogenicity, antibiotic
resistance or novel challenges to host immunity.
The host-pathogen dynamic described for takahe and Campylobacter spp. will be different for different
species of micro-organisms. In my study, this was examined using another enteric organism, Salmonella
spp., which has a greater potential for pathogenic effects. The prevalence of Salmonella spp. in takahe
122
(Chapter 6) was significantly lower than that observed for enteric Campylobacter spp, with population
location being associated with the gastro-intestinal carriage of Salmonella spp. in takahe populations. The
organism was only isolated from takahe resident on one island, and an individual translocated to the breeding
centre (Chapter 6). Salmonella spp. have been historically included in disease screening protocols for takahe
(McInnes et al. 2004), with individuals barred from translocation if they test positive for Salmonella spp.
Potentially this may have prevented the spread of Salmonella spp. throughout the takahe population, thus
explaining the low prevalence observed. However, given the generalist nature of Salmonella spp. and
ubiquity in the environment, it seems unlikely that this explains the spatial patterns observed in this study.
While more work is needed to definitively identify the reservoirs of Salmonella spp., the serotypes isolated,
S. Mississippi and S. houtenae 40;gt were suggestive of the presence of local environmental or animal
reservoirs. In this case, the Salmonella spp. showed no external evidence of causing pathogenic effects on the
takahe. However, Salmonella spp. have caused mortality, enteritis and reproductive disease in birds in New
Zealand (Alley et al. 2002). It is likely that the cases of Salmonella spp. in takahe represented a short term
host-pathogen dynamic where the takahe were spill-over hosts (Chapter 6), Although little is known about
the pathogenicity of Salmonella spp. in takahe, it is nevertheless an example of the potential for pathogen
transmission between nodes of the takahe network (Chapter 6).
The observational design of these studies of bacterial population genomics in a well-described host have a
number of limitations, largely as a result of potential biases introduced by non-random sampling and
confounding between variables. However, they did provide insights into the epidemiology of infectious
organisms in wild fragmented populations. They also highlight the value of working with previously
unidentified host-associated bacterial species and using genomic sequencing. Although it is uncertain
whether an investigation into a commensal organism can be used to predict or prevent an epidemic disease
emergence or transmission, the finding of management associated genomic differences in a commensal
bacterium justifies the need for further investigation of pathogenic organisms within this and other
fragmented ecosystems.
7.2. Disease risks associated with translocations
Disease risks associated with translocations of wildlife were addressed in a seminal publication conducted by
Cunningham et al. (1996). In the last 14 years, significant advances in our ability to detect and monitor
pathogens have improved our epidemiological understanding of pathogen dynamics in natural ecosystems.
123
Here I evaluate how this study has contributed to new perspectives on disease risk associated with current
wildlife management conservation practices.
The role and significance of infectious disease impacts on wildlife remains relatively uncertain. The threat
and impact of disease appears to increase as species move closer to extinction (Heard et al. 2013), justifying
the need for protection against unnecessary pathogen incursions. The IUCN wildlife health specialist group
(IUCN WHSG 2014) and institutes such as the World Organisation for Animal Health (OIE) (OIE 2014)
provide guidance upon such matters relating to disease risks associated with the conservation and
translocation of wildlife (Jakob-Hoff et al. 2014; OIE & IUCN 2014). However, many translocations occur
without sufficient pre- and post-release disease monitoring (Chapter 6 ; Griffith et al. 1993), making it more
difficult to assess the role of infectious disease in programme failure.
Correlations between host and pathogen dynamics are inevitable due to dependency of pathogens on their
host for survival (Tompkins et al. 2011). Thus any changes in host conditions resulting from translocation
may alter this relationship. Although justified for reasons relating to management of small populations,
forced perturbation of population dynamics in endangered populations is likely to increase the risk of
pathogenic incursions (Torchin & Mitchell 2004). Additionally, translocation mediated stress (Teixeira et al.
2007) may induce immunosuppression and increase susceptibility to disease in the species of concern (Kock
et al. 2007; Viggers et al. 1993). Disease can have significant impacts on populations and if an outbreak were
to occur within an endangered species it could take decades to recover, particularly if they have low
reproductive output (Rushmore et al. 2013). Pathogen incursions may not necessarily result in mortality,
instead impacting hosts in subtle ways causing a reduction in fitness, reproduction and changes in behaviour.
Pathogen associated morbidity can be difficult to detect, for example animals infected with Toxoplasma
gondii show no overt symptoms of disease but the protozoan can influence the behaviour of its host,
increasing risk and likelihood of transmission between mice and cats (Berdoy et al. 2000).
Co-evolutionary interactions between hosts and their pathogens can be determined by spatial characteristics
and resulting gene flow (Archie & Ezenwa 2011; Gandon & Nuismer 2009). If infectious organisms are
carried across ecological barriers through the translocation process, allopatric speciation of host-adapted
organisms is a distinct possibility within the new environment (Chapter 4). Alternatively, host management
and the presence of alternative reservoirs may increase opportunities for genetic recombination between
infectious organisms (Chapter 4). For example, antibiotic resistance genes were detected in gram negative
bacteria isolated from faecal samples of the captive brush-tail rock wallaby (Petrogale penicillata) but were
124
not present in the bacteria isolated from the wild populations (Power et al. 2013). The class 1 integrons were
thought to have been acquired from environmental contamination during captivity (Power et al. 2013).
Similarly, the genotype diversity of C. sp. nova 1 carried by takahe populations was associated with host
location (Chapter 4). The functional significance of the microbial diversity in my study was not determined,
but it does demonstrate that translocation has the potential to cause changes to the host’s microbiota. The
unforeseen developments described above would not be mitigated by current practices of targeted pathogen
screening prior to translocation. Most disease surveillance of wildlife translocations is extremely limited with
targeted organisms being identified only to genus or species level. This ignores the possibility of detecting
strain variation, or novel organisms such as the C. sp. nova 1 identified in this study.
If the isolation of wildlife into sub-populations is resulting in biogeographic divergence of micro-organisms
(Chapter 4), a natural question arises whether conservation practices should be less conservative and increase
mixing of sub-populations with the aim of building immunity and resilience for pathogen incursions. A study
of plant-fungal relationships found that highly connected host populations experienced lower incidence of
pathogens due to increased levels of disease resistance (Jousimo et al. 2014). Many infectious organisms
exhibit variation in pathogenicity and can be unpredictable if transmitted to naive hosts; what is a commensal
in one species may be a pathogen in another (Waldenstrom et al. 2010). For instance, Hendra virus is carried
asymptomatically by bats of the genus Pteropus in Australia (Halpin et al. 2011). Incidental viral spill-over
into horses or humans in areas where bat colonies overlap with urban centres (Plowright et al. 2011) results
in disease symptoms and occasionally death in the new host (Daszak et al. 2006; Playford et al. 2010).
Translocation decisions and risk assessments should ideally account for potential reservoirs of pathogens
within a location and the possibility of bidirectional transmission of generalist infectious organisms able to
colonise and transmit between sympatric wildlife. However, the logistical difficulty of accounting for the
increasingly complex interactions of the host and microbiome described in my study make accurate risk
assessment of translocations problematic, especially with the limited budgets available to most conservation
programmes.
Our ability to inform conservation management is hindered by our inability to capture the influence of
complex interactions between hosts and environments and their potential confounding effects on pathogen
diversity. In situ investigations of wildlife epidemics are insightful, however free living animals are
frequently inconspicuous especially if afflicted by illness and the recovery of samples from unhealthy or
deceased individuals may be logistically difficult. Although currently underexploited, non-invasive
epidemiological investigations of host-commensal relationships in wildlife populations are a good proxy for 125
the study of infectious disease epidemiology and as a result are gradually increasing in popularity, e.g. (Bull
et al. 2012; Chapter 3 ; Chapter 4 ; Chapter 5 ; Chiyo et al. 2014). Recent advances in epidemiological tools,
methods and interdisciplinary research are starting to improve our understanding of pathogen dynamics and
may in time improve the accuracy of our forecasting of disease risk associated with translocations of
wildlife.
7.3. Advancing tools for epidemiological investigations of wildlife
Applied disease ecology is a new and emerging concept which is rapidly evolving into an interdisciplinary
field of research. Collaborations between diverse sectors from conservationists through to bioinformaticians
are creating new tools for infectious disease research. Here I discuss how this study has informed new
approaches to epidemiological investigations of host-microbe and by inference host-pathogen relationships
in wildlife.
The development of network models for use on translocation databases is an interesting example of applying
social network methods to conservation management (Grange et al. 2014). In chapter 2, I analysed historic
translocation records to assess takahe sub-population dynamics and identify targets for disease surveillance.
Predictive social network models enable conservation managers to visualise current and historic population
connectivity, and geography to inform the risks associated with translocation of individuals between
fragmented populations.
Epidemiological frameworks have been developed to incorporate genomic data with multivariate models
encompassing both pathogenic and host associated attributes (Chapter 4 ; Kao et al. 2014). Genetic
adaptation of infectious organisms within a relatively short time frame provides an ideal system to explore
the subtle effects of geographic isolation on the evolution and diversity of microbes in fragmented
populations (Chapter 4). Ribosomal multi-locus sequence typing (rMLST) derived from whole genome
sequences is a promising tool for the exploration of genomic differences within and between all bacterial
genera (Jolley et al. 2012). Investigations have shown it is a good proxy for comparisons of the core genome
and requires less computational effort (Chapter 4). rMLST methods provide higher resolution comparisons
than traditional molecular typing (Chapter 3) and output can be incorporated into multivariate analysis of
biotic and abiotic variables acquired from epidemiological investigations (Chapter 4). However, as with
other typing methods reliant on culture of isolates, limitations remain in capturing the full diversity of
microbes present within and between hosts. Although currently costly, genomic typing of pathogens is likely
to be a major contributor to epidemiological investigations in the future. With the invention of portable 126
sequencing tools such as the Oxford MinION (Oxford Nanopore Technologies, Oxford, UK) (Eisenstein
2012), rapid testing of wildlife in field situations may be available in the near future. Understanding the
functional differences that result from changes in microbial genetic diversity will also be important if we are
to reliably predict shifts in important characteristics such as pathogenicity, transmission and antimicrobial
resistance.
7.4. Implications for conservation management
Recommended pre-translocation health and pathogen screening has been founded upon theoretical risk
assessments (Jakob-Hoff et al. 2014; OIE & IUCN 2014). The lack of resolution in current pathogen testing
protocols and empirical research has impeded the formation of robust scientific foundations upon which
costly pathogen screening can be justified. Currently, animals may be translocated with identified or
unknown parasites which may impact health and population reestablishment due to incomplete identification
of pathogens affecting the target species (Sainsbury & Vaughan-Higgins 2012). The development of high
resolution analyses afforded by whole genome sequencing of pathogens, as used in this study, provides a
more incisive method for the investigation of wildlife disease. As the cost of new technologies decreases,
incorporation of these methods into translocation policy is the most beneficial approach to the identification
and mitigation of disease.
Historic movement and infection records have important roles in epidemiological studies and have been used
for retrospective analysis of pathogenic outbreaks in humans (Gardy et al. 2011), domestic animals (Ortiz-
Pelaez et al. 2006) and are now being applied to wildlife. For example, the microbial genomics of
Campylobacter sp. nova 1 may have been shaped by the effects of host isolation, the presence of reservoir
hosts, and heterogeneous population interactions resulting from historic and current population management
(Chapter 4). However, ambiguous movement and health records of translocated individuals limit the ability
to assess the importance of disease in translocation failure (Grange et al. 2014). Informative epidemiological
investigations require accurate recording of movements and maintenance of wildlife health databases for all
government and community controlled population manipulations of wildlife. Similarly, publically archiving
full genome information on current micro-organisms, including commensals carried by these animals, will
allow future studies to document shifts in the microbial dynamics over time.
Captive management and breeding in artificial environments has been the mainstay of population
preservation and augmentation in New Zealand. Intermixing of individuals from sub-populations within a
breeding centre can bring together previously isolated individuals into sympatry. These animals may carry 127
their own endemic infectious organisms which have evolved during a period of isolation into a breeding
facility. This creates an opportunity for a diverse range of potential pathogens to mix in the same
environment (Chapter 4). In order to prevent co-infection and inter-host transmission of pathogens, a period
of quarantine, spelling and decontamination of enclosures is recommended. If a novel pathogen is acquired
during captivity, the process of supplementing wild populations with captive bred individuals derived from
connected ‘hub’ populations could facilitate long range dissemination of potentially novel pathogens
throughout a population network (Grange et al. 2014). High levels of biosecurity and disease screening of all
immigrant and emigrant individuals moving through a breeding facility is highly recommended. However,
disease screening currently relies on a pre-existing knowledge of the micro-organisms which are of
importance to the species in care, balanced against the costs and availability of testing.
In captivity and in some wildlife sanctuaries, the atypical interface between humans, wildlife and livestock
increases interactions and the threat of introduction of exotic pathogens from reservoirs. Although there are
obvious benefits to the targeted use of ‘advocacy individuals’ to promote conservation awareness (e.g.
Sirrocco the Kakapo), vulnerable species maintained in open reserves may be at increased risk of exposure to
novel pathogens via interaction with humans. In an increasingly connected world, international travel
increases the potential for long range pathogen transmission. A notable example is the 2014 reports of people
infected with Ebola virus arriving in uninfected countries including the United States (WHO 2014b) and
Europe (WHO 2014a) after long distance air travel. Interactions between vulnerable wildlife with humans
from around the globe could act as a route for exposure to a complex array of exotic micro-organisms.
Contacts of this nature add another level of complexity and threat when assessing risks of wildlife
translocation.
When considering the choice of locations used for the introduction and maintenance of endangered species,
decisions are rarely based on reducing the threats of pathogen introduction from co-habiting animals. For
example, captive reared Arabian oryx (Oryx leucoryx) succumbed to capripox virus infection transmitted
from domestic sheep grazing along the boundaries of the oryx enclosure (Greth et al. 1992). The installation
of buffer zones around reserves and relative isolation from agricultural sources may aid in reducing the
likelihood of pathogen transmission from exotic sources (Chapter 3), but will increase the costs of
maintaining such reserves. Co-habitation of closely related species within a reserve would not be
recommended since research suggests taxonomically related hosts and those whom share behavioural traits
are more likely to share pathogens (Chapter 5). An investigation into rabies virus in bat species in North
America showed an association between host phylogenetic distance and frequency of cross-species 128
transmission (Streicker et al. 2010). The likelihood of transmission was lower between bat species that were
less closely related (Streicker et al. 2010). Conversely, bringing livestock species directly into contact with
endangered wildlife, such as the sheep that shared the takahe sanctuary of Maud Island (Chapter 5), also
carries risks of exposing the wildlife to novel micro-organisms that have developed undesirable
characteristics such as antibiotic resistance through modern husbandry methods.
Targeting pathways connecting populations through which pathogens can be transmitted has been used as a
means of mitigation of disease spread for centuries. The same principles which applied to the quarantine of
‘typhoid Mary’ in the early 19th Century (Leavitt 1996) are transferable to the mitigation of pathogen spread
between susceptible wildlife populations. Quarantine and biosecurity when moving and introducing
previously isolated animals into new populations are simple but effective measures to reduce risk. In the
hospital setting, education and disinfection measures targeting medical staff that act as bridges between
patients, has helped reduce the transmission of multi-drug resistant bacteria between hospitalised patients
(Allegranzi & Pittet 2009; Pittet et al. 2006). Transferring the same principles to the management of our
endangered wildlife through biosecurity and cleaning of equipment is likely to mitigate the risk of pathogen
spread and introduction.
Although there is no current evidence of management actions causing detrimental effects due to infectious
disease in takahe conservation, the example of a Salmonella spp. positive takahe being moved from a
location known to have Salmonella spp. positive individuals into quarantine at a breeding centre (Chapter 6)
calls into question the efficacy of disease screening recommendations, and more importantly the need for
better communication between researchers, veterinarians and conservation managers. Epidemiological
investigations into the sources of the generalist pathogens, including obtaining prevalence estimates within
each location and testing environmental sources such as food and water sources, can lead to practical
mitigation strategies that reduce the risk of infection in the species of concern.
7.5. Future research directions
Threatened wildlife populations are frequently afflicted by a combination of stressors acting synergistically
(Brook et al. 2008; Heard et al. 2013; Munns 2006), including habitat loss, predation, invasive species and
pathogens. Sampling strategies should account for the mode of transmission of the infectious organism
within and between host populations, reservoirs, and the environment in an attempt to determine transfer
pathways. Provision of genomic epidemiological data to inform transmission models could translate into a
129
more comprehensive understanding of threats or risks posed to a species for use in conservation practice
decisions.
Family and group dynamics of the host can influence microbial carriage in social species (Caillaud et al.
2013). Overlaying pathogen transmission data with population interactions, and host interaction within
populations, was not fully explored in this study but could be approached in a similar manner using
proximity contact data and faecal testing for Campylobacter spp. In a unique study, researchers in New
Zealand are artificially infecting wild brushtail possums (Trichosurus vulpecula) with Mycobacterium bovis,
and are monitoring spread of the organism through social interactions via mark-recapture and proximity
collar network analysis (Rouco et al. 2013).
This exploration of microbial evolution and transmission dynamics in takahe was designed as an
observational cross-sectional study of microbes in the host species. Repeated sampling over time was not
carried out due to the limitations of capturing and sampling wild populations. However, it is recognised that
the relationship between a host and a pathogen can be variable over time. One approach to investigate the
effects of time would be to conduct a longitudinal investigation into host-pathogen dynamics, but this can be
costly in both time and resources. Phylogenetic and coalescent models incorporate a temporal and spatial
component and have been used in the analysis of bacterial genetics derived from wildlife (Girard et al. 2004).
Applying these methods to bacterial genomics is in its infancy, with most reported studies based on viral
genomes (Biek & Real 2010). For example, post hoc phylogenetic and coalescent analysis in combination
with ecological information suggest the novel avian influenza A H7N9 emerged from multiple gene
reassortment events from duck and chicken influenza viruses, suggesting these species may have acted as
intermediate hosts (Liu et al. 2013). Bayesian phylogenetic programmes such as BEAST2 (Bouckaert et al.
2014) could be used to retrospectively analyse the influence of historic management practices on the existing
diversity of infectious organisms within managed populations. For example, it would be interesting to
investigate the temporal component of takahe translocations from the wild to other facilities in the context of
gene recombination between C. sp. nova 1 isolates from those populations.
Microbial CRISPR (clustered regularly interspaced short palindromic repeats) typing is an emerging method
for genomic resolution between bacterial isolates (Cain & Boinett 2013; Kovanen et al. 2014). Bacterial
CRISPRs are DNA loci comprised of short repeat sequences bookmarking non-repetitive spacer DNA
insertions of mobile genetic elements derived from interactions with foreign DNA (Jansen et al. 2002).
CRISPR-cas systems have a role in microbial adaptive immunity and other roles in genetic regulation
130
(Westra et al. 2014). The identification and characterisation of spacer DNA through tools such as
CRISPRTarget (Biswas et al. 2013) and CRISPRDetect (Biswas et al. 2014) can create an effective barcode
of historic interactions between bacterial species, and when combined with host data could provide an
opportunity to record the subtle effects of translocation has upon the gut microbial community.
This study has only just begun to explore the effects of translocation on a single species of micro-organism.
Traditional methods restrict our ability to capture the full diversity of a microbe within a host. Multiple
carriage of Campylobacter spp. was observed within takahe but limitations in detection were apparent when
comparing selective culture and DNA extraction from faecal samples (Chapter 3). Although some
randomisation was used in the selection of isolates, more in-depth studies exploring the microbial genomic
diversity within takahe would help confirm whether the observed biogeographic patterns of C. sp. nova 1
were real not just an artefact of selective sequencing of a limited number of isolates.
Microbial communities occupying a host are the result of complex interactions and niche specialisation. If
this study were to be viewed as an indicator of the effect translocation can have upon a single species, it
would be likely that perturbation of host lifestyle would alter the microbial ecosystem within a host. An
emerging method is to explore differences in microbiota using metagenomic DNA analysis. Comparison of
microbial DNA obtained from skin swabs of co-habiting frogs using metagenomic sequencing of 16S rRNA
genes revealed host species as a predictor of microbial community (McKenzie et al. 2012). Studies in New
Zealand have used 16S rRNA pyrosequencing techniques to assess the gastrointestinal microbiota of
endangered kakapo (Strigops habroptilus) and found differences in microbial communities along the
gastrointestinal tract (Waite et al. 2012), and an association between age and captive rearing on the faecal
microbiota of young birds (Waite et al. 2014). With improvements in faecal extraction methods (Vo &
Jedlicka 2014), a broad scale comparison of microbial communities in translocated populations would be an
interesting avenue to explore.
7.6. Concluding remarks
The evolutionary responses of pathogens to changes in host management and biodiversity are emerging areas
of interest but still require further investigation before they can be applied in a conservation management
framework (Joseph et al. 2013). Research into communicable organisms in translocated populations will
increase our understanding of threats posed to conserved populations, and be a useful traceable system to
inform epidemiological models of infectious disease in natural environments. The case study of the dynamics
of microbes within the takahe hosts presented in this thesis is a preliminary example of the insights 131
fragmented and translocated populations can provide into the epidemiology of epidemics in endangered
species. Epidemiological research in endangered species has been limited for reasons of protection and
minimal disturbance in vulnerable species. This study has been one of the first to demonstrate the benefits of
opportunistic, non-invasive sampling alongside normal management practices. Therefore, I encourage the
integration of non-invasive cross-sectional and longitudinal pathogen investigations into pre- and post-
translocation policies. The use of whole genome sequencing and bioinformatic tools to quantify the genetic
structure of pathogens in wildlife settings is at the leading edge of disease ecology research and is likely to
be more frequently reported in the literature in the future.
132
CHAPTER 8
LITERATURE CITED
133
134
8. Literature Cited
Agudo, R., M. Carrete, M. Alcaide, C. Rico, F. Hiraldo, and J. A. Donázar. 2012. Genetic diversity at neutral
and adaptive loci determines individual fitness in a long-lived territorial bird. Proceedings of the
Royal Society B: Biological Sciences 279:3241-3249.
Allegranzi, B., and D. Pittet. 2009. Role of hand hygiene in healthcare-associated infection prevention.
Journal of Hospital Infection 73:305-315.
Alley, M. R., J. H. Connolly, S. G. Fenwick, G. F. Mackereth, M. J. Leyland, L. E. Rogers, M. Haycock, C.
Nicol, and C. E. M. Reed. 2002. An epidemic of salmonellosis caused by Salmonella Typhimurium
DT160 in wild birds and humans in New Zealand. New Zealand Veterinary Journal 50:170-176.
Alley, M. R., R. A. Fairley, D. G. Martin, L. Howe, and T. Atkinson. 2008. An outbreak of avian malaria in
captive yellowheads/mohua (Mohoua ochrocephala). New Zealand Veterinary Journal 56:247-251.
Almberg, E. S., P. C. Cross, A. P. Dobson, D. W. Smith, and P. J. Hudson. 2012. Parasite invasion following
host reintroduction: a case study of Yellowstone's wolves. Philosophical transactions of the Royal
Society of London. Series B, Biological sciences 367:2840-2851.
Anderson, R. M., and R. M. May. 1986. The invasion, persistence and spread of infectious diseases within
animal and plant communities. Philosophical Transactions of the Royal Society of London. Series B,
Biological sciences 314:533-570.
Antilles, N., A. Sanglas, and M. Cerda-Cuellar. 2013. Free-living waterfowl as a source of zoonotic bacteria
in a dense wild bird population area in northeastern Spain. Transboundary and Emerging
Diseases:12169.
Aragon, T. J. 2012. EpiTools: R package for epidemiologic data and graphics. R package version 0.5-7.
Archie, E. A., and V. O. Ezenwa. 2011. Population genetic structure and history of a generalist parasite
infecting multiple sympatric host species. International Journal for Parasitology 41:89-98.
Archie, E. A., G. Luikart, and V. O. Ezenwa. 2009. Infecting epidemiology with genetics: a new frontier in
disease ecology. Trends in Ecology & Evolution 24:21-30.
Armstrong, D. P., I. Castro, and R. Griffiths. 2007. Using adaptive management to determine requirements
of re-introduced populations: the case of the New Zealand hihi. Journal of Applied Ecology 44:953-
962.
Ashbolt, R., and M. D. Kirk. 2006. Salmonella Mississippi infections in Tasmania: the role of native
Australian animals and untreated drinking water. Epidemiology & Infection 134:1257-1265.
135
Bager, F., and J. Petersen. 1991. Sensitivity and specificity of different methods for the isolation of
Salmonella from pigs. Acta Veterinaria Scandinavica 32:473-481.
Baker, H. G., H. A. Mooney, and J. A. Drake 1986. Ecology of biological invasions of North America and
Hawaii. Springer-Verlag, New York.
Baling, M., W. Ji, B. Gartrell, and D. H. Brunton. 2013. Detection of Salmonella during the translocation of
two endemic New Zealand lizard species within the Hauraki Gulf. New Zealand Journal of Zoology
40:249-254.
Ball, A. 1991. The epidemiology of Salmonella serovars in Tasmania. University of Tasmania.
Ballance, A. 2001. Takahe: The bird that twice came back from the grave. Pages 18-22. The Takahe: 50
Years of Conservation Management and Research. Otago University Press, Dunedin, New Zealand.
Beja-Pereira, A., B. Bricker, S. Chen, C. Almendra, P. J. White, and G. Luikart. 2009. DNA genotyping
suggests that recent brucellosis outbreaks in the Greater Yellowstone area originated from elk.
Journal of Wildlife Diseases 45:1174-1177.
Bellingham, P. J., D. R. Towns, E. K. Cameron, J. J. Davis, D. A. Wardle, J. M. Wilmshurst, and C. P. H.
Mulder. 2010. New Zealand island restoration: seabirds, predators, and the importance of history.
New Zealand Journal of Ecology 34:115-136.
Bengis, R. G., R. A. Kock, and J. Fischer. 2002. Infectious animal diseases: the wildlife livestock interface.
Revue Scientifique Et Technique De L Office International Des Epizooties 21:53-65.
Bennett, J. S., K. A. Jolley, S. G. Earle, C. Corton, S. D. Bentley, J. Parkhill, and M. C. J. Maiden. 2012. A
genomic approach to bacterial taxonomy: an examination and proposed reclassification of species
within the genus Neisseria. Microbiology 158:1570-1580.
Benskin, C. M., K. Wilson, K. Jones, and I. R. Hartley. 2009. Bacterial pathogens in wild birds: a review of
the frequency and effects of infection. Biological reviews of the Cambridge philosophical society
84:349-373.
Benson, D. A., I. Karsch-Mizrachi, D. J. Lipman, J. Ostell, and E. W. Sayers. 2009. GenBank. Nucleic Acids
Research 37:D26-D31.
Berdoy, M., J. P. Webster, and D. W. Macdonald. 2000. Fatal attraction in rats infected with Toxoplasma
gondii. Proceedings of the Royal Society of London. Series B: Biological Sciences 267:1591-1594.
Berger, L., R. Speare, P. Daszak, D. E. Green, A. A. Cunningham, C. L. Goggin, R. Slocombe, M. A. Ragan,
A. D. Hyatt, K. R. McDonald, H. B. Hines, K. R. Lips, G. Marantelli, and H. Parkes. 1998.
Chytridiomycosis causes amphibian mortality associated with population declines in the rain forests
of Australia and Central America. Proceedings of the National Academy of Sciences 95:9031-9036. 136
Besser, T. E., E. F. Cassirer, M. A. Highland, P. Wolff, A. Justice-Allen, K. Mansfield, M. A. Davis, and W.
Foreyt. 2013. Bighorn sheep pneumonia: sorting out the cause of a polymicrobial disease. Preventive
Veterinary Medicine 108:85-93.
Besser, T. E., E. F. Cassirer, C. Yamada, K. A. Potter, C. Herndon, W. J. Foreyt, D. P. Knowles, and S.
Srikumaran. 2012a. Survival of bighorn sheep (Ovis canadensis) commingled with domestic sheep
(Ovis aries) in the absence of Mycoplasma ovipneumoniae. Journal of Wildlife Diseases 48:168-172.
Besser, T. E., M. A. Highland, K. Baker, E. F. Cassirer, N. J. Anderson, J. M. Ramsey, K. Mansfield, D. L.
Bruning, P. Wolff, J. B. Smith, and J. A. Jenks. 2012b. Causes of pneumonia epizootics among
bighorn sheep, western United States, 2008-2010. Emerging Infectious Diseases 18:406-414.
Biek, R., A. O'Hare, D. Wright, T. Mallon, C. McCormick, R. J. Orton, S. McDowell, H. Trewby, R. A.
Skuce, and R. R. Kao. 2012. Whole genome sequencing reveals local transmission patterns of
Mycobacterium bovis in sympatric cattle and badger populations. PLoS Pathogens 8:e1003008.
Biek, R., and L. A. Real. 2010. The landscape genetics of infectious disease emergence and spread.
Molecular Ecology 19:3515-3531.
BirdLife International. 2013. Porphyrio hochstetteri in IUCN 2013, editor. IUCN Red List of Threatened
Species. www.iucnredlist.org.
Biswas, A., P. C. Fineran, and C. M. Brown. 2014. CRISPRDetect: a flexible algorithm to define CRISPR
arrays and CRISPRBank.
Biswas, A., J. N. Gagnon, S. J. J. Brouns, P. C. Fineran, and C. M. Brown. 2013. CRISPRTarget:
bioinformatic prediction and analysis of crRNA targets. RNA Biology 10:817-827.
Blaker, H. 2000. Confidence curves and improved exact confidence intervals for discrete distributions.
Canadian Journal of Statistics 28:783-798.
Boadella, M., C. Gortazar, P. Acevedo, T. Carta, M. P. Martin-Hernando, J. de la Fuente, and J. Vicente.
2011. Six recommendations for improving monitoring of diseases shared with wildlife: examples
regarding mycobacterial infections in Spain. European Journal of Wildlife Research 57:697-706.
Bouckaert, R., J. Heled, D. Kühnert, T. Vaughan, C.-H. Wu, D. Xie, M. A. Suchard, A. Rambaut, and A. J.
Drummond. 2014. BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS
Computational Biology 10:e1003537.
Boyce, W. M., M. E. Weisenberger, M. C. Penedo, and C. K. Johnson. 2011. Wildlife translocation: the
conservation implications of pathogen exposure and genetic heterozygosity. BMC Ecology 11:5.
Branscum, A. J., I. A. Gardner, and W. O. Johnson. 2005. Estimation of diagnostic-test sensitivity and
specificity through Bayesian modeling. Preventive Veterinary Medicine 68:145-163. 137
Brook, B. W., N. S. Sodhi, and C. J. A. Bradshaw. 2008. Synergies among extinction drivers under global
change. Trends in Ecology & Evolution 23:453-460.
Bull, C. M., S. S. Godfrey, and D. M. Gordon. 2012. Social networks and the spread of Salmonella in a
sleepy lizard population. Molecular Ecology 21:4386-4392.
Bull, S. A., V. M. Allen, G. Domingue, F. Jorgensen, J. A. Frost, R. Ure, R. Whyte, D. Tinker, J. E. L.
Corry, J. Gillard-King, and T. J. Humphrey. 2006. Sources of Campylobacter spp. colonizing housed
broiler flocks during rearing. Applied and Environmental Microbiology 72:645-652.
Bunin, J. S., and I. G. Jamieson. 1995. New approaches toward a better understanding of the decline of
takahe (Porphyrio mantelli) in New Zealand. Conservation Biology 9:100-106.
Bunin, J. S., and I. G. Jamieson. 1996. A cross-fostering experiment between the endangered takahe
(Porphyrio mantelli) and its closest relative, the pukeko (P-porphyrio). New Zealand Journal of
Ecology 20:207-213.
Bunin, J. S., I. G. Jamieson, and D. Eason. 1997. Low reproductive success of the endangered takahe
Porphyrio mantelli on offshore island refuges in New Zealand. Ibis 139:144-151.
Cahn, M. L., M. M. Conner, O. J. Schmitz, T. R. Stephenson, J. D. Wehausen, and H. E. Johnson. 2011.
Disease, population viability, and recovery of endangered Sierra Nevada bighorn sheep. The Journal
of Wildlife Management 75:1753-1766.
Caillaud, D., M. E. Craft, and L. A. Meyers. 2013. Epidemiological effects of group size variation in social
species. Journal of the Royal Society Interface 10:1742-5662.
Cain, A. K., and C. J. Boinett. 2013. A CRISPR view of genome sequences. Nature Reviews Microbiology
11:226-226.
Carlson, J. C., A. B. Franklin, D. R. Hyatt, S. E. Pettit, and G. M. Linz. 2011. The role of starlings in the
spread of Salmonella within concentrated animal feeding operations. Journal of Applied Ecology
48:479-486.
Chapter 2. Network analysis of translocated takahe populations to identify disease surveillance targets.
Chapter 3. Using a common commensal bacterium in endangered takahe (Porphyrio hochstetteri), as a
model to explore pathogen dynamics in isolated wildlife populations.
Chapter 4. Wildlife translocation and the evolution and population structure of a host-associated commensal
Campylobacter spp.
Chapter 5. Investigation of vertebrate reservoirs of Campylobacter spp. in an island ecosystem.
Chapter 6. Location specific prevalence of Salmonella spp. in endangered takahe (Porphyrio hochstetteriI).
Chen, H. 2013. VennDiagram: generate high-resolution Venn and Euler plots. R package version 1.6.5. 138
Chiyo, P. I., L. E. Grieneisen, G. Wittemyer, C. J. Moss, P. C. Lee, I. Douglas-Hamilton, and E. A. Archie.
2014. The influence of social structure, habitat, and host traits on the transmission of Escherichia
coli in wild elephants. PLoS One 9:e93408.
Christley, R. M., and N. P. French. 2003. Small-world topology of UK racing: the potential for rapid spread
of infectious agents. Equine Veterinary Journal 35:586-589.
Christley, R. M., G. L. Pinchbeck, R. G. Bowers, D. Clancy, N. P. French, R. Bennett, and J. Turner. 2005.
Infection in social networks: using network analysis to identify high-risk individuals. American
Journal of Epidemiology 162:1024-1031.
Chun-Lung, S. 2014. Betabuster software in http://www.epi.ucdavis.edu/diagnostictests/betabuster.html,
editor.
Clarke, J., K. Warren, M. Calver, P. de Tores, J. Mills, and I. Robertson. 2013. Hematologic and serum
biochemical reference ranges and an assessment of exposure to infectious diseases prior to
translocation of the threatened western ringtail possum (Pseudocheirus occidentalis). Journal of
Wildlife Diseases 49:831-840.
Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian
Journal of Ecology 18:117-143.
Clements, G. M., S. E. Hygnstrom, J. M. Gilsdorf, D. M. Baasch, M. J. Clements, and K. C. Vercauteren.
2011. Movements of white-tailed deer in riparian habitat: implications for infectious diseases. The
Journal of Wildlife Management 75:1436-1442.
Colles, F. M., J. S. Ali, S. K. Sheppard, N. D. McCarthy, and M. C. Maiden. 2011. Campylobacter
populations in wild and domesticated mallard ducks (Anas platyrhynchos). Environmental
Microbiology Reports 3:574-580.
Colles, F. M., N. D. McCarthy, J. C. Howe, C. L. Devereux, A. G. Gosler, and M. C. J. Maiden. 2009.
Dynamics of Campylobacter colonization of a natural host, Sturnus vulgaris (European Starling).
Environmental Microbiology 11:258-267.
Corner, L. A., and P. J. A. Presidente. 1981. Mycobacterium bovis infection in the brush-tailed possum
(Trichosurus vulpecula) .2. Comparison of experimental infections with an Australian cattle strain
and a New Zealand possum strain. Veterinary Microbiology 6:351-366.
Corner, L. A. L., D. U. Pfeiffer, and R. S. Morris. 2003. Social network analysis of Mycobacterium bovis
transmission among captive brushtail possums (Trichosurus vulpecula). Preventive Veterinary
Medicine 59:147-167.
139
Costello, E. K., K. Stagaman, L. Dethlefsen, B. J. M. Bohannan, and D. A. Relman. 2012. The application of
ecological theory toward an understanding of the human microbiome. Science 336:1255-1262.
Cowled, B. D., M. P. Ward, S. W. Laffan, F. Galea, M. G. Garner, A. J. MacDonald, I. Marsh, P. Muellner,
K. Negus, S. Quasim, A. P. Woolnough, and S. D. Sarre. 2012. Integrating survey and molecular
approaches to better understand wildlife disease ecology. PloS One 7:e46310.
Craft, M. E., P. L. Hawthorne, C. Packer, and A. P. Dobson. 2008. Dynamics of a multihost pathogen in a
carnivore community. Journal of Animal Ecology 77:1257-1264.
Craft, M. E., E. Volz, C. Packer, and L. A. Meyers. 2009. Distinguishing epidemic waves from disease
spillover in a wildlife population. Proceedings of the Royal Society B: Biological Sciences
276:1777-1785.
Crouchley, D. 1994. Takahe recovery plan in D. o. Conservation, editor, Wellington, New Zealand.
Cryan, P. M., C. U. Meteyer, J. G. Boyles, and D. S. Blehert. 2013. White-nose syndrome in bats:
illuminating the darkness. BMC Biology 11:47.
Csardi, G., and T. Nepusz. 2006. The igraph software package for complex network research. InterJournal
Complex Systems 1695.
Cummings, K. J., L. D. Warnick, M. A. Davis, K. Eckmann, Y. T. Gröhn, K. Hoelzer, K. MacDonald, T. P.
Root, J. D. Siler, S. M. McGuire, M. Wiedmann, E. M. Wright, S. M. Zansky, and T. E. Besser.
2012. Farm animal contact as risk factor for transmission of bovine-associated Salmonella subtypes.
Emerging Infectious Diseases 18:1929-1936.
Cunningham, A. A. 1996. Disease risks of wildlife translocations. Conservation Biology 10:349-353.
Daszak, P., A. A. Cunningham, and A. D. Hyatt. 2000. Emerging infectious diseases of wildlife - threats to
biodiversity and human health Science 287:443-449.
Daszak, P., R. Plowright, J. H. Epstein, J. Pulliam, S. Abdul Rahman, H. E. Field, C. S. Smith, K. J. Olival,
S. Luby, and K. Halpin. 2006. The emergence of Nipah and Hendra virus: pathogen dynamics across
a wildlife-livestock-human continuum. Pages 186-201 in S. K. Collinge, and C. Ray, editors.
Disease ecology: community structure and pathogen dynamics, Oxford scholarship online.
De Nooy, W., A. Mrvar, and V. Batagelj 2011. Exploratory social network analysis with Pajek. Cambridge
University Press, Cambridge, United Kingdom.
Denis, M., J. Refrégier-Petton, M. J. Laisney, G. Ermel, and G. Salvat. 2001. Campylobacter contamination
in French chicken production from farm to consumers. Use of a PCR assay for detection and
identification of Campylobacter jejuni and Camp. coli. Journal of Applied Microbiology 91:255-
267. 140
Dobson, A. 2004. Population dynamics of pathogens with multiple host species. American Naturalist
164:S64-S78.
Drewe, J. A., K. T. D. Eames, J. R. Madden, and G. P. Pearce. 2011. Integrating contact network structure
into tuberculosis epidemiology in meerkats in South Africa: implications for control. Preventive
Veterinary Medicine 101:113-120.
Drummond, A., B. Ashton, M. Cheung, J. Heled, M. Kearse, and R. Moir. 2013. Geneious Pro in Biomatters,
editor.
Dube, C., C. Ribble, D. Kelton, and B. McNab. 2009. A review of network analysis terminology and its
application to foot-and-mouth disease modelling and policy development. Transboundary and
Emerging Diseases 56:73-85.
Eason, D. K., and M. Willans. 2001. Captive rearing: a management tool for the recovery of the endangered
takahe. Pages 80-95 in W. G. Lee, and I. G. Jamieson, editors. The Takahe: 50 Years of
Conservation Management and Research. Otago University Press, Dunedin, New Zealand.
Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic
Acids Research 32:1792-1797.
EFSA. 2013. The European Union summary report on trends and sources of zoonoses, zoonotic agents and
food-borne outbreaks in 2011. Page 250. EFSA Journal European Centre for Disease Prevention and
Control.
Eisenstein, M. 2012. Oxford Nanopore announcement sets sequencing sector abuzz. Nat Biotech 30:295-296.
ESR. 2010. Public health surveillance: non human Salmonella isolates.
Ewen, J. G., D. P. Armstrong, R. Empson, S. Jack, T. Makan, K. McInnes, K. A. Parker, K. Richardson, and
M. Alley. 2012. Parasite management in translocations: lessons from a threatened New Zealand bird.
Oryx 46:446-456.
Ewen, J. G., R. Thorogood, C. Nicol, D. P. Armstrong, and M. Alley. 2007. Salmonella typhimurium in hihi,
New Zealand. Emerging Infectious Diseases 13:788-790.
Excoffier, L., and H. E. Lischer. 2010. Arlequin suite ver 3.5: a new series of programs to perform
population genetics analyses under Linux and Windows. Molecular Ecology Resources 10:564-567.
Faria, N. R., M. A. Suchard, A. Rambaut, D. G. Streicker, and P. Lemey. 2013. Simultaneously
reconstructing viral cross-species transmission history and identifying the underlying constraints.
Philosophical Transactions of the Royal Society B-Biological Sciences 368.
Fischer, J., and D. B. Lindenmayer. 2000. An assessment of the published results of animal relocations.
Biological Conservation 96:1-11. 141
Fisher, M. C., D. A. Henk, C. J. Briggs, J. S. Brownstein, L. C. Madoff, S. L. McCraw, and S. J. Gurr. 2012.
Emerging fungal threats to animal, plant and ecosystem health. Nature 484:186-194.
Foley, J., D. Clifford, K. Castle, P. Cryan, and R. S. Ostfeld. 2011. Investigating and managing the rapid
emergence of white nose syndrome, a novel, fatal, infectious disease of hibernating bats.
Conservation Biology 25:223-231.
Foley, J. A., R. Defries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin, M. T. Coe, G. C.
Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C. J. Kucharik, C. Monfreda, J. A.
Patz, I. C. Prentice, N. Ramankutty, and P. K. Snyder. 2005. Global consequences of land use.
Science 309:570-574.
French, N. P., A. Midwinter, B. Holland, J. Collins-Emerson, R. Pattison, F. Colles, and P. Carter. 2009.
Molecular epidemiology of Campylobacter jejuni isolates from wild-bird fecal material in children's
playgrounds. Applied and Environmental Microbiology 75:779-783.
French, N. P., S. Yu, P. Biggs, B. Holland, P. Fearnhead, B. Binney, A. Fox, D. Grove-White, J. W. Leigh,
W. Miller, P. Muellner, and P. Carter. 2014. Evolution of Campylobacter Species in New Zealand.
Pages 221-240 in S. K. Sheppard, editor. Campylobacter Ecology and Evolution. Caister Academic
Press, Norfolk, UK.
Frick, W. F., J. F. Pollock, A. C. Hicks, K. E. Langwig, D. S. Reynolds, G. G. Turner, C. M. Butchkoski, and
T. H. Kunz. 2010. An emerging disease causes regional population collapse of a common North
American bat species. Science 329:679-682.
Gandon, S., A. Buckling, E. Decaestecker, and T. Day. 2008. Host-parasite coevolution and patterns of
adaptation across time and space. Journal of Evolutionary Biology 21:1861-1866.
Gandon, S., and S. L. Nuismer. 2009. Interactions between genetic drift, gene flow, and selection mosaics
drive parasite local adaptation. The American Naturalist 173:212-224.
Gardy, J. L., J. C. Johnston, S. J. H. Sui, V. J. Cook, L. Shah, E. Brodkin, S. Rempel, R. Moore, Y. Zhao,
and R. Holt. 2011. Whole genome sequencing and social network analysis of a tuberculosis
outbreak. New England Journal of Medicine 364:730-739.
Gartrell, B. D., M. R. Alley, H. Mack, J. Donald, K. McInnes, and P. Jansen. 2005. Erysipelas in the
critically endangered kakapo (Strigops habroptilus). Avian Pathology 34:383-387.
Gartrell, B. D., E. Jillings, B. A. Adlington, H. Mack, and N. J. Nelson. 2006. Health screening for a
translocation of captive-reared tuatara (Sphenodon punctatus) to an island refuge. New Zealand
Veterinary Journal 54:344-349.
142
Gilbert, M. J., M. Kik, A. J. Timmerman, T. T. Severs, J. G. Kusters, B. Duim, and J. A. Wagenaar. 2014.
Occurrence, diversity, and host association of intestinal Campylobacter, Arcobacter, and
Helicobacter in reptiles. PLoS One 9:e101599.
Girard, J. M., D. M. Wagner, A. J. Vogler, C. Keys, C. J. Allender, L. C. Drickamer, and P. Keim. 2004.
Differential plague-transmission dynamics determine Yersinia pestis population genetic structure on
local, regional, and global scales. Proceedings of the National Academy of Sciences 101:8408-8413.
Godfrey, S. S. 2013. Networks and the ecology of parasite transmission: a framework for wildlife
parasitology. International Journal for Parasitology: Parasites and Wildlife 2:235-245.
Goering, R. V., and C. A. Mims 2013. Mims' medical microbiology. Saunders, Edinburgh.
Gonzalez, I., K. A. Grant, P. T. Richardson, S. F. Park, and M. D. Collins. 1997. Specific identification of
the enteropathogens Campylobacter jejuni and Campylobacter coli by using a PCR test based on the
ceuE gene encoding a putative virulence determinant. Journal of Clinical Microbiology 35:759-763.
Grange, Z. L., M. Van Andel, N. P. French, and B. D. Gartrell. 2014. Network analysis of translocated
takahe populations to identify disease surveillance targets. Conservation Biology 28:518-528.
Greenacre, M. 2007. Correspondence analysis in practice. Chapman & Hall/CRC, Florida, USA.
Greth, A., J. M. Gourreau, M. Vassart, Nguyenbavy, M. Wyers, and P. C. Lefevre. 1992. Capripoxvirus
disease in an Arabian oryx (Oryx-Leucoryx) from Saudi Arabia. Journal of Wildlife Diseases
28:295-300.
Griekspoor, P., F. M. Colles, N. D. McCarthy, P. M. Hansbro, C. Ashhurst-Smith, B. Olsen, D. Hasselquist,
M. C. Maiden, and J. Waldenstrom. 2013. Marked host specificity and lack of phylogeographic
population structure of Campylobacter jejuni in wild birds. Molecular Ecology 22:1463-1472.
Griekspoor, P., B. Olsen, and J. Waldenström. 2009. Campylobacter jejuni in penguins, Antarctica.
Emerging Infectious Diseases 14:847-849.
Griffith, B., J. M. Scott, J. W. Carpenter, and C. Reed. 1989. Translocation as a species conservation tool:
status and strategy. Science 245:477-480.
Griffith, B., J. M. Scott, J. W. Carpenter, and C. Reed. 1993. Animal translocations and potential disease
transmission. Journal of Zoo and Wildlife Medicine 24:231-236.
Grimont, P. A. D., and F. X. Weill. 2007. Antigenic formulae of the Salmonella serovars in WHO, editor.
Collaborating centre for reference and research on Salmonella. Institut Pasteur, Paris, France.
Grueber, C. E., R. J. Laws, S. Nakagawa, and I. G. Jamieson. 2010. Inbreeding depression accumulation
across life-history stages of the endangered takahe. Conservation Biology 24:1617-1625.
143
Guivier, E., M. Galan, Y. Chaval, A. Xuereb, A. Ribas Salvador, M. L. Poulle, L. Voutilainen, H. Henttonen,
N. Charbonnel, and J. F. Cosson. 2011. Landscape genetics highlights the role of bank vole
metapopulation dynamics in the epidemiology of Puumala hantavirus. Molecular Ecology 20:3569-
3583.
Hall, A. J., and E. K. Saito. 2008. Avian wildlife mortality events due to salmonellosis in the United States,
1985-2004. Journal of Wildlife Diseases 44:585-593.
Halpin, K., A. D. Hyatt, R. Fogarty, D. Middleton, J. Bingham, J. H. Epstein, S. A. Rahman, T. Hughes, C.
Smith, H. E. Field, and P. Daszak. 2011. Pteropid bats are confirmed as the reservoir hosts of
henipaviruses: a comprehensive experimental study of virus transmission. The American Journal of
Tropical Medicine and Hygiene 85:946-951.
Hand, B. K., S. A. Cushman, E. L. Landguth, and J. Lucotch. 2014. Assessing multi-taxa sensitivity to the
human footprint, habitat fragmentation and loss by exploring alternative scenarios of dispersal ability
and population size: a simulation approach. Biodiversity and Conservation 23:2761-2779.
Hartley, M., and E. Gill. 2010. Assessment and mitigation processes for disease risks associated with wildlife
management and conservation interventions. Veterinary Record 166:487-490.
Haydon, D. T., S. Cleaveland, L. H. Taylor, and M. K. Laurenson. 2002. Identifying reservoirs of infection:
a conceptual and practical challenge. Emerging Infectious Diseases 8:1468-1473.
Heard, M. J., K. F. Smith, K. J. Ripp, M. Berger, J. Chen, J. Dittmeier, M. Goter, S. T. McGarvey, and E.
Ryan. 2013. The threat of disease increases as species move toward extinction. Conservation
Biology 27:1378-1388.
Heather, B. D., and H. A. Robertson 2006. The field guide to the birds of New Zealand. Vicking, Auckland.
Hegg, D., G. Greaves, J. M. Maxwell, D. I. MacKenzie, and I. G. Jamieson. 2012. Demography of takahe
(Porphyrio hochstetteri) in Fiordland: environmental factors and management affect survival and
breeding success. New Zealand Journal of Ecology 36:75-89.
Hess, G. 1996. Disease in Metapopulation Models: Implications for Conservation. Ecology 77:1617-1632.
Hess, G. R. 1994. Conservation corridors and contagious disease - a cautionary note. Conservation Biology
8:256-262.
Hilty, J. A., C. Brooks, E. Heaton, and A. M. Merenlender. 2006. Forecasting the effect of land-use change
on native and non-native mammalian predator distributions. Biodiversity and Conservation 15:2853-
2871.
Hirsch, B. T., S. Prange, S. A. Hauver, and S. D. Gehrt. 2013. Raccoon social networks and the potential for
disease transmission. PloS One 8:e75830. 144
Huang, S., O. R. P. Bininda-Emonds, P. R. Stephens, J. L. Gittleman, and S. Altizer. 2014. Phylogenetically
related and ecologically similar carnivores harbour similar parasite assemblages. Journal of Animal
Ecology 83:671-680.
Hughes, L. A., M. Bennett, P. Coffey, J. Elliott, T. R. Jones, R. C. Jones, A. Lahuerta-Marin, A. H.
Leatherbarrow, K. McNiffe, D. Norman, N. J. Williams, and J. Chantrey. 2009. Molecular
epidemiology and characterization of Campylobacter spp. isolated from wild bird populations in
northern England. Applied and Environmental Microbiology 75:3007-3015.
Huson, D. H., and D. Bryant. 2006. Application of phylogenetic networks in evolutionary studies. Molecular
Biology and Evolution 23:254-267.
Husson, F., J. Josse, S. Le, and J. Mazet. 2013. FactoMineR: multivariate exploratory data analysis and data
mining with R. Version 1.25.
IUCN WHSG. 2014. IUCN Wildlife Health Specialist Group, http://www.iucn-whsg.org/about.
Ivanek, R., J. Österberg, R. Gautam, and S. Sternberg Lewerin. 2012. Salmonella fecal shedding and immune
responses are dose- and serotype- dependent in pigs. PloS One 7:e34660.
Iveson, J. B., G. R. Shellam, S. D. Bradshaw, D. W. Smith, J. S. Mackenzie, and R. G. Mofflin. 2009.
Salmonella infections in Antarctic fauna and island populations of wildlife exposed to human
activities in coastal areas of Australia. Epidemiology & Infection 137:858-870.
Jakob-Hoff, R., S. C. MacDiarmid, C. Lees, P. S. Miller, D. Travis, R. A. Kock, International Office of
Epizootics., and International Union for Conservation of Nature., editors. 2014. Manual of
procedures for wildlife disease risk analysis. World Organisation for Animal Health.
Jamieson, I. G., and C. J. Ryan. 1999. Causes of low reproductve success of translocated takahe (Porphorio
hochstetteri) on predator free islands. Science for Conservation. Department of Conservation,
Wellington, New Zealand.
Jamieson, I. G., and C. J. Ryan. 2001. Island takahe: Closure of the debate over the merits of introducing
Fiordland takahe to predator-free islands. Pages 96-113 in I. G. Jamieson, and W. G. Lee, editors.
The Takahe: Fifty Years of Conservation Management and Research. Otago University Press,
Dunedin, New Zealand.
Jamieson, I. G., G. P. Wallis, and J. V. Briskie. 2006. Inbreeding and endangered species management: is
New Zealand out of step with the rest of the world? Conservation Biology 20:38-47.
Jansen, R., J. D. A. Embden, W. Gaastra, and L. M. Schouls. 2002. Identification of genes that are associated
with DNA repeats in prokaryotes. Molecular Microbiology 43:1565-1575.
145
Jeger, M. J., M. Pautasso, O. Holdenrieder, and M. W. Shaw. 2007. Modelling disease spread and control in
networks: implications for plant sciences. The New Phytologist 174:279-297.
Johnson, P. T. J., J. R. Rohr, J. T. Hoverman, E. Kellermanns, J. Bowerman, and K. B. Lunde. 2012. Living
fast and dying of infection: host life history drives interspecific variation in infection and disease
risk. Ecology Letters 15:235-242.
Jolley, K. A., C. M. Bliss, J. S. Bennett, H. B. Bratcher, C. Brehony, F. M. Colles, H. Wimalarathna, O. B.
Harrison, S. K. Sheppard, A. J. Cody, and M. C. Maiden. 2012. Ribosomal multilocus sequence
typing: universal characterization of bacteria from domain to strain. Microbiology 158:1005-1015.
Jones, K. 2001. Campylobacters in water, sewage and the environment. Journal of Applied Microbiology
90:68S-79S.
Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D. Balk, J. L. Gittleman, and P. Daszak. 2008. Global
trends in emerging infectious diseases. Nature 451:990-993.
Joseph, M. B., J. R. Mihaljevic, A. L. Arellano, J. G. Kueneman, D. L. Preston, P. C. Cross, and P. T. J.
Johnson. 2013. Taming wildlife disease: bridging the gap between science and management. Journal
of Applied Ecology 50:702-712.
Jousimo, J., A. J. M. Tack, O. Ovaskainen, T. Mononen, H. Susi, C. Tollenaere, and A.-L. Laine. 2014.
Ecological and evolutionary effects of fragmentation on infectious disease dynamics. Science
344:1289-1293.
Kao, R. R., D. T. Haydon, S. J. Lycett, and P. R. Murcia. 2014. Supersize me: how whole-genome
sequencing and big data are transforming epidemiology. Trends in Microbiology 22:282-291.
Keller, J. I., and W. G. Shriver. 2013. Prevalence of three Campylobacter species, C. Jejuni, C. Coli, and C.
Lari, using multilocus sequence typing in wild birds of the Mid-Atlantic region, USA. Journal of
Wildlife Diseases 50:31-41.
Keller, J. I., W. G. Shriver, J. Waldenström, P. Griekspoor, and B. Olsen. 2011. Prevalence of
Campylobacter in Wild Birds of the Mid-Atlantic Region, USA. Journal of wildlife diseases 47:750-
754.
Kerr, P. J. 2012. Myxomatosis in Australia and Europe: a model for emerging infectious diseases. Antiviral
Research 93:387-415.
Kilpatrick, A. M., and S. Altizer. 2010. Disease ecology. Nature Education Knowledge 2:13.
Kittl, S., G. Heckel, B. M. Korczak, and P. Kuhnert. 2013. Source attribution of human Campylobacter
isolates by MLST and Fla-typing and association of genotypes with Quinolone resistance. PloS One
8:e81796. 146
Kock, R. A., P. S. Soorae, and O. B. Mohammed. 2007. Role of veterinarians in re-introductions.
International Zoo Yearbook 41:24-37.
Kolde, R. 2013. pheatmap: pretty heatmaps.
Kovanen, S. M., R. I. Kivistö, M. Rossi, and M. L. Hänninen. 2014. A combination of MLST and CRISPR
typing reveals dominant Campylobacter jejuni types in organically farmed laying hens. Journal of
Applied Microbiology 117:249-257.
Kupferschmidt, K. 2014. Estimating the Ebola epidemic. Science 345:1108-1108.
Kwan, P. S., M. Barrigas, F. J. Bolton, N. P. French, P. Gowland, R. Kemp, H. Leatherbarrow, M. Upton,
and A. J. Fox. 2008. Molecular epidemiology of Campylobacter jejuni populations in dairy cattle,
wildlife, and the environment in a farmland area. Applied and Environmental Microbiology
74:5130-5138.
Lawson, B., T. Howard, J. K. Kirkwood, S. K. Macgregor, M. Perkins, R. A. Robinson, L. R. Ward, and A.
A. Cunningham. 2010. Epidemiology of salmonellosis in garden birds in England and Wales, 1993
to 2003. EcoHealth 7:294-306.
Leavitt, J. W. 1996. Typhoid Mary: captive to the public's health. Beacon Press.
Lee, W. G. 2001. Fifty years of takahe conservation, research and management: what have we learnt? Pages
49-60 in W. G. Lee, editor. The Takahe: 50 years of Conservation Management and Research. Otago
University Press, Dunedin, New Zealand.
Li, L., C. J. Stoeckert, and D. S. Roos. 2003. OrthoMCL: identification of ortholog groups for Eukaryotic
genomes. Genome Research 13:2178-2189.
Linton, D., A. J. Lawson, R. J. Owen, and J. Stanley. 1997. PCR detection, identification to species level,
and fingerprinting of Campylobacter jejuni and Campylobacter coli direct from diarrheic samples.
Journal of Clinical Microbiology 35:2568-2572.
Liu, D., W. Shi, Y. Shi, D. Wang, H. Xiao, W. Li, Y. Bi, Y. Wu, X. Li, J. Yan, W. Liu, G. Zhao, W. Yang,
Y. Wang, J. Ma, Y. Shu, F. Lei, and G. F. Gao. 2013. Origin and diversity of novel avian influenza
A H7N9 viruses causing human infection: phylogenetic, structural, and coalescent analyses. The
Lancet 381:1926-1932.
Lloyd-Smith, J. O., S. J. Schreiber, P. E. Kopp, and W. M. Getz. 2005. Superspreading and the effect of
individual variation on disease emergence. Nature 438:355-359.
Maiden, M. C. J., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. Zhou, K. Zurth,
D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 1998. Multilocus sequence typing: a
147
portable approach to the identification of clones within populations of pathogenic microorganisms.
Proceedings of the National Academy of Sciences 95:3140-3145.
Marin, C., S. Ingresa-Capaccioni, S. González-Bodi, F. Marco-Jiménez, and S. Vega. 2013. Free-living
turtles are a reservoir for Salmonella but not for Campylobacter. PloS one 8:e72350.
Marin, C., M.-D. Palomeque, F. Marco-Jiménez, and S. Vega. 2014. Wild griffon vultures (Gyps fulvus) as a
source of Salmonella and Campylobacter in eastern Spain. PloS one 9:e94191.
Mathews, F., D. Moro, R. Strachan, M. Gelling, and N. Buller. 2006. Health surveillance in wildlife
reintroductions. Biological Conservation 131:338-347.
Mawdsley, J. R., R. O'Malley, and D. S. Ojima. 2009. A review of climate-change adaptation strategies for
wildlife management and biodiversity conservation. Conservation Biology 23:1080-1089.
Maxwell, J. M., and I. G. Jamieson. 1997. Survival and recruitment of captive reared and wild-reared Takahe
in Fiordland, New Zealand. Conservation Biology 11:683-691.
McCallum, H., N. Barlow, and J. Hone. 2001. How should pathogen transmission be modelled? Trends in
Ecology & Evolution 16:295-300.
McInnes, K., P. Cromarty, and J. Ombler. 2004. Standard operating procedure for the health management of
terrestrial vertebrate species protected under the wildlife act. Department of Conservation,
Wellington, , NZ.
McKenzie, V. J., R. M. Bowers, N. Fierer, R. Knight, and C. L. Lauber. 2012. Co-habiting amphibian
species harbor unique skin bacterial communities in wild populations. ISME J 6:588-596.
McLelland, J. M., B. D. Gartrell, and W. D. Roe. 2011. A retrospective study of post-mortem examination
findings in takahe (Porphyrio hochstetteri). New Zealand Veterinary Journal 59:160-165.
Meagher, M., and M. E. Meyer. 1994. On the origin of brucellosis in bison of Yellowstone National Park - a
review. Conservation Biology 8:645-653.
Middleton, D. M. R. L., A. C. La Flamme, B. D. Gartrell, and N. J. Nelson. 2014. Reptile reservoirs and
seasonal variation in the environmental presence of Salmonella in an island ecosystem, Stephens
Island, New Zealand. Journal of Wildlife Diseases 50:655-659.
Miskelly, C. M., J. E. Dowding, G. P. Elliott, R. A. Hitchmough, R. P. Powlesland, H. A. Robertson, P. M.
Sagar, R. P. Scofield, and G. A. Taylor. 2008. Conservation status of New Zealand birds. Notornis
55:117-135.
Mohan, V., M. Stevenson, J. Marshall, P. Fearnhead, B. R. Holland, G. Hotter, and N. P. French. 2013.
Campylobacter jejuni colonization and population structure in urban populations of ducks and
starlings in New Zealand. MicrobiologyOpen 2:659-673. 148
Morner, T., D. L. Obendorf, M. Artois, and M. H. Woodford. 2002. Surveillance and monitoring of wildlife
diseases. Revue Scientifique Et Technique De L Office International Des Epizooties 21:67-76.
Morris, R. S., and D. U. Pfeiffer. 1995. Directions and issues in bovine tuberculosis epidemiology and
control in New Zealand. New Zealand Veterinary Journal 43:256-265.
Mughini Gras, L., J. H. Smid, J. A. Wagenaar, A. G. d. Boer, A. H. Havelaar, I. H. M. Friesema, N. P.
French, L. Busani, and W. van Pelt. 2012. Risk factors for campylobacteriosis of chicken, ruminant,
and environmental origin: a combined case-control and source attribution analysis. PloS one
7:e42599-e42599.
Mullner, P., J. M. Collins-Emerson, A. C. Midwinter, P. Carter, S. E. Spencer, P. van der Logt, S. Hathaway,
and N. P. French. 2010. Molecular epidemiology of Campylobacter jejuni in a geographically
isolated country with a uniquely structured poultry industry. Applied and Environmental
Microbiology 76:2145-2154.
Mullner, P., S. E. F. Spencer, D. J. Wilson, G. Jones, A. D. Noble, A. C. Midwinter, J. M. Collins-Emerson,
P. Carter, S. Hathaway, and N. P. French. 2009. Assigning the source of human campylobacteriosis
in New Zealand: a comparative genetic and epidemiological approach. Infection, Genetics and
Evolution 9:1311-1319.
Munnoch, S. A., K. Ward, S. Sheridan, G. J. Fitzsimmons, C. T. Shadbolt, J. P. Piispanen, Q. Wang, T. J.
Ward, T. L. M. Worgan, C. Oxenford, J. A. Musto, J. McAnulty, and D. N. Durrheim. 2009. A
multi-state outbreak of Salmonella Saintpaul in Australia associated with cantaloupe consumption.
Epidemiology & Infection 137:367-374.
Munns, W. R. 2006. Assessing risks to wildlife populations from multiple stressors: overview of the problem
and research needs. Ecology and Society 11:23.
Murphy, C., C. Carroll, and K. N. Jordan. 2006. Environmental survival mechanisms of the foodborne
pathogen Campylobacter jejuni. Journal of Applied Microbiology 100:623-632.
Murray, N. 2004. Handbook on import risk analysis for animals and animal products, Paris, France.
Newman, M. E. J. 2010. Networks : an introduction. Oxford University Press, Oxford ; New York.
Nugent, G., E. J. Whitford, I. Yockney, M. Perry, D. M. Tompkins, N. Holtslag, and M. L. Cross. 2013.
Percutaneous interdigital injection of Mycobacterium bovis as a model for tuberculous lesion
development in wild brushtail possums (Trichosurus vulpecula). Journal of Comparative Pathology
148:33-42.
Nunn, C. L., and A. T. W. Dokey. 2006. Ranging patterns and parasitism in primates. Biology Letters 2:351-
354. 149
Nunn, C. L., P. H. Thrall, and P. M. Kappeler. 2014. Shared resources and disease dynamics in spatially
structured populations. Ecological Modelling 272:198-207.
Nunn, C. L., P. H. Thrall, F. H. Leendertz, and C. Boesch. 2011. The spread of fecally transmitted parasites
in socially-structured populations. PloS One 6:e21677.
Obrien, S. J., and J. F. Evermann. 1988. Interactive influence of infectious disease and genetic diversity in
natural populations. Trends in Ecology & Evolution 3:254-259.
OIE. 2011. Annual report 2011. Office International des Epizooties, Paris.
OIE. 2014. http://www.oie.int/about-us/history/.
OIE, and IUCN 2014. Guidelines for wildlife disease risk analysis. World Organisation for Animal Health.
Olival, K. J., R. L. Hoguet, and P. Daszak. 2013. Linking the historical roots of environmental conservation
with human and wildlife health. EcoHealth 10:224-227.
Opsahl, T. 2009. Structure and evolution of weighted networks. University of London (Queen Mary
College), London, United Kingdom.
Opsahl, T., F. Agneessens, and J. Skvoretz. 2010. Node centrality in weighted networks: generalizing degree
and shortest paths. Social Networks 32:245-251.
Opsahl, T., V. Colizza, P. Panzarasa, and J. J. Ramasco. 2008. Prominence and control: the weighted rich-
club effect. Physical Review Letters 101:168702.
Opsahl, T., and P. Panzarasa. 2009. Clustering in weighted networks. Social Networks 31:155-163.
Orr, M. 1997. Salmonella Brandenburg in a takahe. Page 14 in W. S. o. NZVA, editor. Kokako.
Ortiz-Pelaez, A., D. U. Pfeiffer, R. J. Soares-Magalhaes, and F. J. Guitian. 2006. Use of social network
analysis to characterize the pattern of animal movements in the initial phases of the 2001 foot and
mouth disease (FMD) epidemic in the UK. Preventive Veterinary Medicine 76:40-55.
Palacios, G., L. J. Lowenstine, M. R. Cranfield, K. V. K. Gilardi, L. Spelman, M. Lukasik-Braum, J. F.
Kinani, A. Mudakikwa, E. Nyirakaragire, A. V. Bussetti, N. Savji, S. Hutchison, M. Egholm, and W.
I. Lipkin. 2011. Human metapneumovirus infection in wild mountain gorillas, Rwanda. Emerging
Infectious Diseases 17:711-713.
Parker, K. A. 2008. Translocations: providing outcomes for wildlife, resource managers, scientists, and the
human community. Restoration Ecology 16:204-209.
Parsons, S. K., C. M. Bull, and D. M. Gordon. 2010. Low prevalence of Salmonella enterica in Australian
wildlife. Environmental Microbiology Reports 2:657-659.
Peakall, R., and P. E. Smouse. 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for
teaching and research—an update. Bioinformatics 28:2537-2539. 150
Pennycott, T. W., H. A. Mather, G. Bennett, and G. Foster. 2010. Salmonellosis in garden birds in Scotland,
1995 to 2008: geographic region, Salmonella enterica phage type and bird species. Veterinary
Record 166:419-421.
Pikula, J., H. Bandouchova, L. Novotný, C. U. Meteyer, J. Zukal, N. R. Irwin, J. Zima, and N. Martínková.
2012. Histopathology confirms white-nose syndrome in bats in Europe. Journal of Wildlife Diseases
48:207-211.
Pittet, D., B. Allegranzi, H. Sax, S. Dharan, C. L. Pessoa-Silva, L. Donaldson, and J. M. Boyce. 2006.
Evidence-based model for hand transmission during patient care and the role of improved practices.
The Lancet Infectious Diseases 6:641-652.
Playford, E. G., B. McCall, G. Smith, V. Slinko, G. Allen, I. Smith, F. Moore, C. Taylor, Y. H. Kung, and H.
Field. 2010. Human hendra virus encephalitis associated with equine outbreak, Australia, 2008.
Emerging Infectious Diseases 16:219-223.
Plowright, R. K., P. Foley, H. E. Field, A. P. Dobson, J. E. Foley, P. Eby, and P. Daszak. 2011. Urban
habituation, ecological connectivity and epidemic dampening: The emergence of Hendra virus from
flying foxes (Pteropus spp.). Proceedings of the Royal Society B: Biological Sciences 278:3703-
3712.
Plowright, R. K., S. H. Sokolow, M. E. Gorman, P. Daszak, and J. E. Foley. 2008. Causal inference in
disease ecology: investigating ecological drivers of disease emergence. Frontiers in Ecology and the
Environment 6:420-429.
Porphyre, T., J. McKenzie, and M. A. Stevenson. 2011. Contact patterns as a risk factor for bovine
tuberculosis infection in a free-living adult brushtail possum Trichosurus vulpecula population.
Preventive Veterinary Medicine 100:221-230.
Power, M. L., S. Emery, and M. R. Gillings. 2013. Into the wild: Dissemination of antibiotic resistance
determinants via a species recovery program. PloS One 8:e63017.
Pracy, L. T. 1974. Introduction and liberation of the opossum into New Zealand. Page 28. New Zealand
forest service information series
Prenter, J., C. MacNeil, J. T. A. Dick, and A. M. Dunn. 2004. Roles of parasites in animal invasions. Trends
in Ecology & Evolution 19:385-390.
R Core Team. 2013. R: a language and environment for statistical computing. R Foundation for Statistical
Computing, Vienna, Austria.
151
Rachowicz, L. J., R. A. Knapp, J. A. Morgan, M. J. Stice, V. T. Vredenburg, J. M. Parker, and C. J. Briggs.
2006. Emerging infectious disease as a proximate cause of amphibian mass mortality. Ecology
87:1671-1683.
Real, L. A., and R. Biek. 2007. Spatial dynamics and genetics of infectious diseases on heterogeneous
landscapes. Journal of the Royal Society Interface 4:935-948.
Rose, K. 2000. Takahe (Porphyrio hochstetteri) Health Investigation Project. Department of Conservation,
Wellington, NZ.
Rouco, C., K. Richardson, N. P. French, B. Buddle, C. Jewell, and D. M. Tompkins. 2013. Determinants of
possum home range characteristics, and their implications for TB transmission. Page 18 in J.
Hufschmid, I. Beveridge, C. Death, and P. Whiteley, editors. Wildlife Disease Association
Australasian Section Annual Conference, Grampians National Park, Vic, Australia.
Rushmore, J., D. Caillaud, L. Matamba, R. M. Stumpf, S. P. Borgatti, and S. Altizer. 2013. Social network
analysis of wild chimpanzees provides insights for predicting infectious disease risk. The Journal of
Animal Ecology 82:976-986.
Rushton, S. P., P. W. W. Lurz, J. Gurnell, and R. Fuller. 2000. Modelling the spatial dynamics of
parapoxvirus disease in red and grey squirrels: a possible cause of the decline in the red squirrel in
the UK? Journal of Applied Ecology 37:997-1012.
Rushton, S. P., P. W. W. Lurz, J. Gurnell, P. Nettleton, C. Bruemmer, M. D. F. Shirley, and A. W.
Sainsbury. 2006. Disease threats posed by alien species: the role of a poxvirus in the decline of the
native red squirrel in Britain. Epidemiology and Infection 134:521-533.
Rwego, I. B., G. Isabirye-Basuta, T. R. Gillespie, and T. L. Goldberg. 2008. Gastrointestinal bacterial
transmission among humans, mountain gorillas, and livestock in Bwindi Impenetrable National Park,
Uganda. Conservation Biology 22:1600-1607.
Sainsbury, A. W., R. Deaville, B. Lawson, W. A. Cooley, S. S. J. Farelly, M. J. Stack, P. Duff, C. J.
McInnes, J. Gurnell, P. H. Russell, S. P. Rushton, D. U. Pfeiffer, P. Nettleton, and P. W. W. Lurz.
2008. Poxviral disease in red squirrels Sciurus vulgaris in the UK: spatial and temporal trends of an
emerging threat. EcoHealth 5:305-316.
Sainsbury, A. W., and R. J. Vaughan-Higgins. 2012. Analyzing disease risks associated with translocations.
Conservation Biology 26:442-452.
Saito, E. K., and A. J. Hall. 2008. Avian wildlife mortality events due to salmonellosis in the United States,
1985-2004. Journal of Wildlife Diseases 44:585-593.
152
Sayers, E. W., T. Barrett, D. A. Benson, S. H. Bryant, K. Canese, V. Chetvernin, D. M. Church, M.
DiCuccio, R. Edgar, S. Federhen, M. Feolo, L. Y. Geer, W. Helmberg, Y. Kapustin, D. Landsman,
D. J. Lipman, T. L. Madden, D. R. Maglott, V. Miller, I. Mizrachi, J. Ostell, K. D. Pruitt, G. D.
Schuler, E. Sequeira, S. T. Sherry, M. Shumway, K. Sirotkin, A. Souvorov, G. Starchenko, T. A.
Tatusova, L. Wagner, E. Yaschenko, and J. Ye. 2009. Database resources of the National Center for
Biotechnology Information. Nucleic Acids Research 37:D5-D15.
Seemann, T. 2014. Prokka: rapid prokaryotic genome annotation. Bioinformatics.
Sheppard, S. K., F. Colles, J. Richardson, A. J. Cody, R. Elson, A. Lawson, G. Brick, R. Meldrum, C. L.
Little, R. J. Owen, M. C. Maiden, and N. D. McCarthy. 2010. Host association of Campylobacter
genotypes transcends geographic variation. Applied and Environmental Microbiology 76:5269-5277.
Sheppard, S. K., F. M. Colles, N. D. McCarthy, N. J. C. Strachan, I. D. Ogden, K. J. Forbes, J. F. Dallas, and
M. C. J. Maiden. 2011. Niche segregation and genetic structure of Campylobacter jejuni populations
from wild and agricultural host species. Molecular Ecology 20:3484-3490.
Sherley, G., I. Stringer, and G. Parrish. 2010. Summary of native bat, reptile, amphibian and terrestrial
invertebrate translocations in New Zealand. Pages 1-39 in D. o. Conservation, editor. Science for
Conservation.
Silvy, N. J. 2012. The wildlife techniques manual. Johns Hopkins University Press, Baltimore, USA.
Simberloff, D., and J. Cox. 1987. Consequences and costs of conservation corridors. Conservation Biology
1:63-71.
Sippy, R., C. M. J. Sandoval-Green, O. Sahin, P. Plummer, W. S. Fairbanks, Q. Zhang, and J. A. Blanchong.
2012. Occurrence and molecular analysis of Campylobacter in wildlife on livestock farms.
Veterinary Microbiology 157:369-375.
Skerratt, L., L. Berger, R. Speare, S. Cashins, K. McDonald, A. D. Phillott, H. B. Hines, and N. Kenyon.
2007. Spread of chytridiomycosis has caused the rapid global decline and extinction of frogs.
EcoHealth 4:125-134.
Smith, K. F., K. Acevedo-Whitehouse, and A. B. Pedersen. 2009. The role of infectious diseases in
biological conservation. Animal Conservation 12:1-12.
Smith, K. F., D. F. Sax, and K. D. Lafferty. 2006. Evidence for the role of infectious disease in species
extinction and endangerment. Conservation Biology 20:1349-1357.
Snyder, B., J. Thilsted, B. Burgess, and M. Richard. 1985. Pigeon herpesvirus mortalities in foster reared
Mauritius pink pigeons. Pages 60-70. American Association of Zoo Veterinarians, Atlanta, Georgia,
USA. 153
Soorae, P. S. 2008. The IUCN position statement on translocation of living organisms. Pages 266-284.
Global reintroduction perspectives: re-introduction case-studies from around the globe. International
Union for the Conservation of Nature / Species Survival Commission, Re-introduction Specialist
Group, Abu Dhabi, United Arab Emirates.
Sousa, V., and J. Hey. 2013. Understanding the origin of species with genome-scale data: modelling gene
flow. Nature Reviews Genetics 14:404-414.
Stevenson, M. 2014. epiR: an R package for the analysis of epidemiological data.
Streicker, D. G., A. S. Turmelle, M. J. Vonhof, I. V. Kuzmin, G. F. McCracken, and C. E. Rupprecht. 2010.
Host phylogeny constrains cross-species emergence and establishment of rabies virus in bats.
Science 329:676-679.
Stucki, U., J. Frey, J. Nicolet, and A. P. Burnens. 1995. Identification of Campylobacter jejuni on the basis
of a species-specific gene that encodes a membrane protein. Journal of Clinical Microbiolgy 33 855-
859.
Sullivan, L. L., B. L. Johnson, L. A. Brudvig, and N. M. Haddad. 2011. Can dispersal mode predict corridor
effects on plant parasites? Ecology 92:1559-1564.
Tear, T. H., J. M. Scott, P. H. Hayward, and B. Griffith. 1993. Status and prospects for success of the
endangered species act - a look at recovery plans. Science 262:976-977.
Teixeira, C. P., C. S. de Azevedo, M. Mendl, C. F. Cipreste, and R. J. Young. 2007. Revisiting translocation
and reintroduction programmes: the importance of considering stress. Animal Behaviour 73:1-13.
Tennyson, A., and P. Martinson 2006. Extinct birds of New Zealand. Te Papa Press, Wellington, New
Zealand.
Thompson, R. C., A. J. Lymbery, and A. Smith. 2010. Parasites, emerging disease and wildlife conservation.
International Journal for Parasitology 40:1163-1170.
Tompkins, D. M., A. M. Dunn, M. J. Smith, and S. Telfer. 2011. Wildlife diseases: from individuals to
ecosystems. Journal of Animal Ecology 80:19-38.
Torchin, M. E., K. D. Lafferty, A. P. Dobson, V. J. McKenzie, and A. M. Kuris. 2003. Introduced species
and their missing parasites. Nature 421:628-630.
Torchin, M. E., and C. E. Mitchell. 2004. Parasites, pathogens, and invasions by plants and animals.
Frontiers in Ecology and the Environment 2:183-190.
Trewick, S. A. 1996. Morphology and evolution of two takahe: flightless rails of New Zealand. Journal of
Zoology 238:221-237.
154
Trewick, S. A., and T. H. Worthy. 2001. Origins and prehistoric ecology of takahe based on morphometric,
molecular, and fossil data. Pages 31-48 in W. G. Lee, and I. G. Jamieson, editors. The Takahe: 50
years of Conservation Management and Research. Otago University Press, Dunedin, New Zealand.
Van Immerseel, F., J. De Buck, F. Pasmans, L. Bohez, F. Boyen, F. Haesebrouck, and R. Ducatelle. 2004.
Intermittent long-term shedding and induction of carrier birds after infection of chickens early
posthatch with a low or high dose of Salmonella Enteritidis. Poultry Science 83:1911-1916.
VanderWaal, K. L., E. R. Atwill, S. Hooper, K. Buckle, and B. McCowan. 2013a. Network structure and
prevalence of Cryptosporidium in Belding’s ground squirrels. Behavioral Ecology and Sociobiology
67:1951-1959.
VanderWaal, K. L., E. R. Atwill, L. A. Isbell, and B. McCowan. 2013b. Linking social and pathogen
transmission networks using microbial genetics in giraffe (Giraffa camelopardalis). Journal of
Animal Ecology 83:406-414.
VanderWaal, K. L., E. R. Atwill, L. A. Isbell, and B. McCowan. 2014. Quantifying microbe transmission
networks for wild and domestic ungulates in Kenya. Biological Conservation 169:136-146.
VanWormer, E., M. A. Miller, P. A. Conrad, M. E. Grigg, D. Rejmanek, T. E. Carpenter, and J. A. K. Mazet.
2014. Using molecular epidemiology to track Toxoplasma gondii from terrestrial carnivores to
marine hosts: implications for public health and conservation. PloS Neglected Tropical Diseases
8:e2852.
Verbrugghe, E., F. Boyen, A. Van Parys, K. Van Deun, S. Croubels, A. Thompson, N. Shearer, B. Leyman,
F. Haesebrouck, and F. Pasmans. 2011. Stress induced Salmonella Typhimurium recrudescence in
pigs coincides with cortisol induced increased intracellular proliferation in macrophages. Veterinary
Research 42:118.
Viggers, K. L., D. B. Lindenmayer, and D. M. Spratt. 1993. The importance of disease in reintroduction
programmes. Wildlife Research 20:687-698.
Vittor, A. Y., R. H. Gilman, J. M. Tielsch, G. Glass, T. Shields, W. Sanchez-Lozano, V. V. Pinedo, and J. A.
Patz. 2005. The effect of deforestation on the human-biting rate and larval breeding sites of
Anopheles darlingi, the primary vector of Falciparum malaria in the Peruvian Amazon. American
Journal of Tropical Medicine and Hygiene 73:150-150.
Vlahović, K., B. Matica, I. Bata, M. Pavlak, Ž. Pavičić, M. Popović, S. Nejedli, and A. Dovč. 2004.
Campylobacter, Salmonella and Chlamydia in free-living birds of Croatia. European Journal of
Wildlife Research 50:127-132.
155
Vo, A. T., and J. A. Jedlicka. 2014. Protocols for metagenomic DNA extraction and Illumina amplicon
library preparation for faecal and swab samples. Mol Ecol Resour 14:1183–1197.
Waite, D. W., P. Deines, and M. W. Taylor. 2012. Gut microbiome of the critically endangered New Zealand
parrot, the kakapo (Strigops habroptilus). PLoS One 7:e35803.
Waite, D. W., D. K. Eason, and M. W. Taylor. 2014. Influence of hand-rearing and bird age on the faecal
microbiota of the critically endangered kakapo. Applied and Environmental Microbiology 80:4650.
Waldenstrom, J., D. Axelsson-Olsson, B. Olsen, D. Hasselquist, P. Griekspoor, L. Jansson, S. Teneberg, L.
Svensson, and P. Ellstrom. 2010. Campylobacter jejuni colonization in wild birds: results from an
infection experiment. Plos One 5:e9082.
Waldenstrom, J., T. Broman, I. Carlsson, D. Hasselquist, R. P. Achterberg, J. A. Wagenaar, and B. Olsen.
2002. Prevalence of Campylobacter jejuni, Campylobacter lari, and Campylobacter coli in different
ecological guilds and taxa of migrating birds. Applied and environmental microbiology 68:5911-
5917.
Waldenstrom, J., and P. Griekspoor. 2014. Ecology and host associations of Campylobacter in wild birds.
Pages 265-284 in S. Sheppard , and G. Méric, editors. Campylobacter Ecology and Evolution.
Horizon Scientific Press, Norfolk, UK.
Warner, R. E. 1968. Role of introduced diseases in extinction of endemic Hawaiian avifauna. Condor
70:101-120.
Watts, D. J., and S. H. Strogatz. 1998. Collective dynamics of small-world networks. Nature 393:440-442.
Weeks, A. R., C. M. Sgro, A. G. Young, R. Frankham, N. J. Mitchell, K. A. Miller, M. Byrne, D. J. Coates,
M. D. Eldridge, P. Sunnucks, M. F. Breed, E. A. James, and A. A. Hoffmann. 2011. Assessing the
benefits and risks of translocations in changing environments: a genetic perspective. Evolutionary
applications 4:709-725.
Westra, E. R., A. Buckling, and P. C. Fineran. 2014. CRISPR-Cas systems: beyond adaptive immunity. Nat
Rev Micro 12:317-326.
Wey, T., D. T. Blumstein, W. Shen, and F. Jordán. 2008. Social network analysis of animal behaviour: a
promising tool for the study of sociality. Animal Behaviour 75:333-344.
WHO. 2014a. Ebola virus disease – Spain in G. A. a. R. (GAR), editor. Disease outbreak news
WHO. 2014b. Ebola virus disease – United States of America in G. A. a. R. (GAR), editor. Disease outbreak
news
Wickes, C., D. Crouchley, and J. M. Maxwell. 2009. Takahe (Porphorio hochstetteri) recovery plan 2007 -
2012. Threatened Species Recovery Plan 61. Department of Conservation, Wellington, NZ. 156
Wickham, H. 2009. ggplot2: elegant graphics for data analysis. Springer, New York.
Williams, J. M., E. V. Lonsdorf, M. L. Wilson, J. Schumacher-Stankey, J. Goodall, and A. E. Pusey. 2008.
Causes of death in the Kasekela chimpanzees of Gombe National Park, Tanzania. American Journal
of Primatology 70:766-777.
Winfield, M. D., and E. A. Groisman. 2003. Role of nonhost environments in the lifestyles of Salmonella
and Escherichia coli. Applied and Environmental Microbiology 69:3687-3694.
Wobeser, G. A. 2006. Essentials of disease in wild animals. Blackwell, Ames, USA.
Woodford, M. H. 1993. International disease implications for wildlife translocation. Journal of Zoo and
Wildlife Medicine 24:265-270.
Woodford, M. H. 2009. Veterinary aspects of ecological monitoring: the natural history of emerging
infectious diseases of humans, domestic animals and wildlife. Tropical Animal Health and
Production 41:1023-1033.
Woodford, M. H., and P. B. Rossiter. 1993. Disease risks associated with wildlife translocation projects.
Revue scientifique et technique of the Office International des Epizooties:115-135.
Woodford, M. H., and P. B. Rossiter. 1994. Disease risks associated with wildlife translocation projects.
Pages 178-200 in P. J. S. Olney, G. M. Mace, and A. T. C. Feistner, editors. Creative Conservation.
Springer Netherlands.
Woolhouse, M. E. J., C. Dye, J. F. Etard, T. Smith, J. D. Charlwood, G. P. Garnett, P. Hagan, J. L. K. Hii, P.
D. Ndhlovu, R. J. Quinnell, C. H. Watts, S. K. Chandiwana, and R. M. Anderson. 1997.
Heterogeneities in the transmission of infectious agents: implications for the design of control
programs. Proceedings of the National Academy of Sciences 94:338-342.
Woolhouse, M. E. J., D. J. Shaw, L. Matthews, W. C. Liu, D. J. Mellor, and M. R. Thomas. 2005.
Epidemiological implications of the contact network structure for cattle farms and the 20-80 rule.
Biology Letters 1:350-352.
Zerbino, D. R., and E. Birney. 2008. Velvet: algorithms for de novo short read assembly using de Bruijn
graphs. Genome Research 18:821-829.
157
158
CHAPTER 9
APPENDIX
159
160
9. Appendix
9.2. Chapter 2 supplementary information
Appendix 9.2-1 Map of New Zealand (including region boundaries) and key demonstrating the location of takahe reserves and treatment centres involved in this network analysis. Node H is not shown as it is a private island not under control of the NZ Department of Conservation.
161
Year From location From letter To location To letter 2007 Burwood Bush breeding centre A Maud Island D 2007 Burwood Bush breeding centre A Tiritiri Matangi Island J 2007 Mana Island C Wildbase Hopsital G 2007 Wildbase Hopsital G Maud Island D 2007 Tiritiri Matangi Island J Burwood Bush breeding centre A 2007 Tiritiri Matangi Island J Burwood Bush breeding centre A 2007 Tiritiri Matangi Island J Burwood Bush breeding centre A 2007 Tiritiri Matangi Island J Mana Island C 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Burwood Bush breeding centre A Murchison Mountains Q 2007 Tiritiri Matangi Island J Murchison Mountains Q 2007 Tiritiri Matangi Island J Murchison Mountains Q 2007 Tiritiri Matangi Island J Murchison Mountains Q 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2007 Murchison Mountains Q Burwood Bush breeding centre A 2008 Burwood Bush breeding centre A Mana Island C 2008 Burwood Bush breeding centre A Mana Island C 2008 Burwood Bush breeding centre A Mana Island C 2008 Burwood Bush breeding centre A Wildbase Hopsital G 2008 Burwood Bush breeding centre A Wildbase Hopsital G 2008 Kapiti Island B Burwood Bush breeding centre A 2008 Mana Island C Burwood Bush breeding centre A 2008 Mana Island C Burwood Bush breeding centre A 2008 Mana Island C Pukaha Mt Bruce F 2008 Mana Island C Pukaha Mt Bruce F 2008 Maud Island D Burwood Bush breeding centre A 2008 Pukaha Mt Bruce F Wildbase Hopsital G 2008 Wildbase Hopsital G Mana Island C 2008 Wildbase Hopsital G Pukaha Mt Bruce F
162
Year From From letter To To letter 2008 Tiritiri Matangi Island J Burwood Bush breeding centre A 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Burwood Bush breeding centre A Murchison Mountains Q 2008 Mana Island C Murchison Mountains Q 2008 Tiritiri Matangi Island J Murchison Mountains Q 2008 Maud Island D Murchison Mountains Q 2008 Mana Island C Murchison Mountains Q 2008 Maud Island D Murchison Mountains Q 2008 Murchison Mountains Q Burwood Bush breeding centre A 2008 Murchison Mountains Q Burwood Bush breeding centre A 2008 Murchison Mountains Q Burwood Bush breeding centre A 2009 Burwood Bush breeding centre A Maud Island D 2009 Burwood Bush breeding centre A Secretary Island M 2009 Burwood Bush breeding centre A Secretary Island M 2009 Kapiti Island B Burwood Bush breeding centre A 2009 Kapiti Island B Wildbase Hopsital G 2009 Kapiti Island B Wildbase Hopsital G 2009 Mana Island C Burwood Bush breeding centre A 2009 Mana Island C Burwood Bush breeding centre A 2009 Mana Island C Burwood Bush breeding centre A 2009 Mana Island C Maungatautari Island reseve E 2009 Maud Island D Maungatautari Island reseve E 2009 Wildbase Hopsital G Burwood Bush breeding centre A 2009 Wildbase Hopsital G Kapiti Island B 2009 Wildbase Hopsital G Te Anau Wildlife Reserve I 2009 Private island H Peacock Springs Wildlife Park L 2009 Private island H Peacock Springs Wildlife Park L 2009 Private island H Peacock Springs Wildlife Park L 2009 Te Anau Wilidlife Reserve I Wildbase Hopsital G 2009 Tiritiri Matangi Island J Burwood Bush breeding centre A 2009 Tiritiri Matangi Island J Burwood Bush breeding centre A 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Burwood Bush breeding centre A Murchison Mountains Q
163
Year From From letter To To letter 2009 Burwood Bush breeding centre A Murchison Mountains Q 2009 Mana Island C Murchison Mountains Q 2009 Tiritiri Matangi Island J Murchison Mountains Q 2009 Kapiti Island B Murchison Mountains Q 2009 Tiritiri Matangi Island J Murchison Mountains Q 2009 Mana Island C Murchison Mountains Q 2009 Murchison Mountains Q Burwood Bush breeding centre A 2009 Murchison Mountains Q Burwood Bush breeding centre A 2009 Murchison Mountains Q Burwood Bush breeding centre A 2009 Murchison Mountains Q Burwood Bush breeding centre A 2010 Burwood Bush breeding centre A Kapiti Island B 2010 Burwood Bush breeding centre A Kapiti Island B 2010 Burwood Bush breeding centre A Maud Island D 2010 Burwood Bush breeding centre A Wildbase Hopsital G 2010 Burwood Bush breeding centre A Wildbase Hopsital G 2010 Burwood Bush breeding centre A Secretary Island M 2010 Burwood Bush breeding centre A Secretary Island M 2010 Burwood Bush breeding centre A Secretary Island M 2010 Burwood Bush breeding centre A Willowbank Reserve O 2010 Burwood Bush breeding centre A Willowbank Reserve O 2010 Kapiti Island B Burwood Bush breeding centre A 2010 Kapiti Island B Burwood Bush breeding centre A 2010 Kapiti Island B Burwood Bush breeding centre A 2010 Kapiti Island B Burwood Bush breeding centre A 2010 Kapiti Island B Tiritri Matangi J 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Burwood Bush breeding centre A 2010 Mana Island C Wildbase Hopsital G 2010 Wildbase Hopsital G Burwood Bush breeding centre A 2010 Wildbase Hopsital G Te Anau Wildlife Reserve I 2010 Private island H Secretary Island M 2010 Te Anau Wilidlife Reserve I Wildbase Hopsital G 2010 Te Anau Wilidlife Reserve I Wildbase Hopsital G 2010 Tiritiri Matangi Island J Burwood Bush breeding centre A 2010 Tiritiri Matangi Island J Burwood Bush breeding centre A 2011 Burwood Bush breeding centre A Private island H 2011 Burwood Bush breeding centre A Secretary Island M 2011 Burwood Bush breeding centre A Secretary Island M
164
Year From From letter To To letter 2011 Burwood Bush breeding centre A Secretary Island M 2011 Kapiti Island B Wildbase Hopsital G 2011 Mana Island C Wildbase Hopsital G 2011 Mana Island C Motutapu Island K 2011 Mana Island C Motutapu Island K 2011 Mana Island C Karori Wildlife Sanctuary P 2011 Mana Island C Karori Wildlife Sanctuary P 2011 Maud Island D Burwood Bush breeding centre A 2011 Maud Island D Wildbase Hopsital G 2011 Maud Island D Wildbase Hopsital G 2011 Maungatautari Reserve E Wildbase Hopsital G 2011 Maungatautari Reserve E Wildbase Hopsital G 2011 Maungatautari Reserve E Motutapu Island K 2011 Wildbase Hopsital G Burwood Bush breeding centre A 2011 Wildbase Hopsital G Kapiti Island B 2011 Wildbase Hopsital G Mana Island C 2011 Wildbase Hopsital G Mana Island C 2011 Wildbase Hopsital G Maud Island D 2011 Wildbase Hopsital G Maungatautari Island reseve E 2011 Wildbase Hopsital G Maungatautari Island reseve E 2011 Te Anau Wilidlife Reserve I Wellington Zoo N 2011 Tiritiri Matangi Island J Motutapu Island K Appendix 9.2-2 Dataset includes records of individual movements of takahe (Porphyrio hochstetteri) between locations in New Zealand from 2007 to 2011.
165
2007 2007 to 2008 2007 to 2009 2007 to 2010 2007 to 2011 Location ki
in kiout Bi ki
in kiout Bi ki
in kiout Bi ki
in kiout Bi ki
in kiout Bi
Burwood Bush breeding centre 4 3 4 10 12 21 14 16 42 17 21 52 18 23 78 Kapiti Island - - - 0 1 0 1 4 0 2 6 0 3 7 0 Mana Island 3 1 4 4 5 5 4 8 11 4 11 8 5 14 19 Maud Island 3 0 0 2 3 6 2 4 0 3 4 4 3 6 0 Maungatautari reserve - - - - - - 2 0 0 2 0 0 3 2 0 Pukaha Mt Bruce - - - 2 1 0 2 1 0 2 1 0 2 1 0 Wildbase Hospital 1 2 2 3 3 5 6 6 30 8 7 26 11 10 61 Private Island - - - - - - 0 2 0 0 3 0 1 3 10 Te Anau wildlife reserve - - - - - - 1 1 0 1 2 0 1 3 9 Tiritiri Matangi Island 1 7 4 1 5 0 1 6 0 2 7 0 2 8 0 Motutapu Island - - - - - - - - - - - - 3 0 0 Peacock springs wildlife park - - - - - - 2 0 0 2 0 0 2 0 0 Secretary Island - - - - - - 1 0 0 3 0 0 4 0 0 Wellington Zoo - - - - - - - - - - - - 1 0 0 Willowbank reserve - - - - - - - - - 1 0 0 1 0 0 Zealandia / Karori sanctuary - - - - - - - - - - - - 1 0 0 Murchison Mountains 3 3 1 11 3 0 15 4 0 15 4 0 15 4 0 Key player A J ACD Q A A Q A A A A A A A A
Appendix 9.2-3 Node level measures, in degree (kiin), out degree (ki
out), and betweenness (Bi), for takahe (Porphyrio hochstetteri) networks used for selection of key players according to highest node measure per year.
166
9.3. Chapter 3 supplementary information
167
168
169
170
Appendix 9.3-1 R script for latent class analysis in diagnostic testing
171
Model 1a Model 1b Model 2a Model 2b Model 3a Model 3b Se1 91.4 (84.6-96.5) 52.4 (30.8-83.0) 92.5 (86.5-96.9) 56.4 (31.8-87.1) 97.8 (93.8-99.7) 96.1 (86.8-99.8) Sp1 88.3 (83.1-92.8) 85.6 (68.9-99.3) 90.7 (85.9-94.5) 96.4 (89.4-99.8) 88.6 (82.8-93.4) 40.1 (14.6-86.5) Se2 62.9 (52.2-73.5) 34.2 (15.2-65.7) 65.3 (54.4-75.7) 30.6 (14.8-52.2) 69.1 (61.9-75.9) 78.5 (60.8-98.0) Sp2 95.7 (91.2-99.1) 94.0 (82.8-99.7) 97.7 (94.8-99.6) 97.7 (92.7-100.0) 97.8 (93.8-99.7) 77.3 (46.0-99.0) Pr1 16.4 (4.3-33.5) 22.6 (0.01-63.0) 30.6 (12.4-49.7) 60.6 (29.0-94.3) 81.7 (67.2-92.2) 68.1 (28.6-93.3) Pr2 40.7 (21.7-59.5) 70.8 (27.6-98.4) 6.3 (1.2-14.8) 0.05 (0.00-15.7) 81.2 (67.8-91.0) 69.2 (31.4-91.7) Pr3 20.4 (8.8-33.8) 28.6 (6.2-63.1) 24.7 (10.6-38.9) 42.8 (19.9-75.6) 81.2 (68.1-92.4) 63.9 (24.0-91.2) P (Pr1>Pr2) 0.02 0.01 0.001 0 0.471 0.489 P (Pr1>Pr3) 0.334 0.367 0.292 0.15 0.472 0.371 P (Pr2> Pr3) 0.035 0.031 0.009 0 0.482 0.361
Appendix 9.3-2 True sensitivity and specificity of faecal DNA / PCR (Se1/ Sp1) and culture / PCR (Se2/ Sp2) for the detection of Campylobacter jejuni (model 1), Campylobacter coli (model 2) and Campylobacter sp. nova 1 (model 3), with true prevalence in the three populations; breeding (Pr1), wild (Pr2) and insurance (Pr3) was estimated by conditionally independent Bayesian latent class analysis with (a) informative and (b) non-informative priors, β (1,1). P is a Bayesian statistical probability estimating differences in prevalence (Pr) between populations where values close to 0 and 1 indicate potential significant differences.
172
Appendix 9.3-3 Multiple correspondence analysis (MCA) plots demonstrating a) covariate relationships between 118 takahe (Porphyrio hochstetteri) where coordinates of the first two dimensions (Dim.1 and Dim.2) explain most of the variation in the data represented. Observations are presented as grey dots on the plot, with density curves illustrating zones where individual observations overlap. Variable categories are assigned by name with subcategories of a variable being colour coordinated. Hierarchical clustering of MCA by b) population management and c) location betweenness.
173
Variable Level Coefficient (SE) p-value ChiSq p-value Sex Intercept Female - - 0.11 * Male 0.62 (0.39) 0.12 Age Intercept Adult - - 0.76 Juvenile -0.14 (0.45) 0.76 Rearing Intercept Wild - - 0.97 Foster -0.24 (1.25) 0.85 Puppet -0.07 (0.42) 0.87 Nest site Intercept North Island - - 0.53 South Island 0.28 (0.44) 0.53 Nest site Intercept Insurance - - 0.05 * Breeding 1.16 (0.54) 0.03 Wild 0.92 (0.46) 0.05 Location Intercept North Island - - 0.09 * South Island 0.83 (0.51) 0.11 Location Intercept Insurance - - <0.01 * Breeding 0.36 (0.54) 0.50 Wild 1.65 (0.49) <0.01 Location Intercept Offshore - - 0.01 * Mainland 1.22 (0.50) 0.02 Translocation Intercept No - - 0.84 Yes -0.08 (0.41) 0.84 In degree Intercept High - - 0.01 * Low -1.08 (0.44) 0.01 Out degree Intercept High - - 0.36 Low 0.37 (0.41) 0.37 Betweenness Intercept High - - 0.18 * Low 0.69 (0.45) 0.12 Medium -0.08 (0.65) 0.90 Location Intercept Close to farms - - 0.18 * Remote from farms 0.54 (0.41) 0.18 Nest site Intercept Close to farms - - 0.50 Remote from farms -0.29 (0.43) 0.50 Sampling period Intercept Period 1 (Mar-Apr 2012) - - 0.02 * Period 2 (Aug-Nov 2012) -0.56 (0.77) 0.58 Period 3 (Feb-Apr 2013) -2.75 (0.46) 0.01 *variable was included in a multivariate model
Appendix 9.3-4 Table of Campylobacter jejuni univariate analyses
174
Variable Level Coefficient (SE) p-value ChiSq p-value Sex Intercept Female - - 0.56 Male 0.39 (0.68) 0.57 Age Intercept Adult - - 0.59 Juvenile 0.38 (0.71) 0.60 Rearing Intercept Wild - - 0.66 Foster 14.50 (1385.38) 0.99 Puppet 0.30 (0.70) 0.67 Nest site Intercept North Island - - 0.19 * South Island 0.84 (0.63) 0.18 Nest site Intercept Insurance - - 0.34 Breeding 1.42 (1.11) 0.20 Wild 0.41 (0.65) 0.53 Location Intercept North Island - - 0.68 South Island -0.33 (0.81) 0.69 Location Intercept Insurance - - 0.93 Breeding 0.14 (0.80) 0.86 Wild -0.14 (0.71) 0.84 Location Intercept Offshore - - 0.92 Mainland -0.07 (0.70) 0.92 Translocation Intercept No - - 0.47 Yes -0.46 (0.62) 0.46 In degree Intercept High - - 0.96 Low 0.03 (0.65) 0.97 Out degree Intercept High - - 0.97 Low -0.03 (0.65) 0.97 Betweenness Intercept High - - 0.65 Low -0.03 (0.74) 0.97 Medium -0.73 (0.88) 0.41 Location Intercept Close to farms - - 0.81 Remote from farms -0.15 (0.65) 0.81 Nest site Intercept Close to farms - - 0.11 * Remote from farms -1.42 (1.07) 0.18 Sampling period Intercept Period 1 (Mar-Apr 2012) - - 0.52 Period 2 (Aug-Nov 2012) -0.01 (2306.1) 0.99 Period 3 (Feb-Apr 2013) -0.39 (0.86) 0.70 *variable was included in a multivariate model
Appendix 9.3-5 Table of Campylobacter sp. nova 1 univariate analyses
175
Variable Level Coefficient (SE) p-value ChiSq p-value Sex Intercept Female - - 0.25 Male -0.51 (0.44) 0.25 Age Intercept Adult - - <0.01 * Juvenile 1.39 (0.47) <0.01 Rearing Intercept Wild - - 0.10 * Foster 2.16 (1.26) 0.09 Puppet 0.69 (0.46) 0.14 Nest site Intercept North Island - - 0.06 * South Island -0.88 (0.46) 0.06 Nest site Intercept Insurance - - 0.03 * Breeding -0.12 (0.54) 0.83 Wild -1.30 (0.55) 0.02 Location Intercept North Island - - 0.29 South Island -0.54 (0.28) 0.28 Location Intercept Insurance - - <0.01 * Breeding 0.61 (0.49) 0.21 Wild -2.91 (1.07) 0.01 Location Intercept Offshore - - 0.50 Mainland 0.34 (0.52) 0.51 Translocation Intercept No - - 0.65 Yes 0.21 (0.46) 0.65 In degree Intercept High - - 0.26 Low 0.51 (0.45) 0.25 Out degree Intercept High - - <0.01 * Low -1.30 (0.45) <0.01 Betweenness Intercept High - - <0.01 * Low -1.61 (0.50) <0.01 Medium -1.02 (0.67) 0.13 Location Intercept Close to farms - - <0.01 * Remote from farms -2.23 (0.50) <0.01 Nest site Intercept Close to farms - - 0.02 * Remote from farms -1.1 (0.46) 0.02 Sampling period Intercept Period 1 (Mar-Apr 2012) - - 0.25 Period 2 (Aug-Nov 2012) 0.64 (0.78) 0.52 Period 3 (Feb-Apr 2013) -1.35 (0.52) 0.18 *variable was included in a multivariate model
Appendix 9.3-6 Table of Campylobacter coli univariate analyses
176
Variable Level Coefficient (SE) p-value ChiSq p-value Sex Intercept Female - - 0.44 Male -0.68 (0.89) 0.45 Age Intercept Adult - - 0.63 Juvenile 0.39 (0.80) 0.63 Rearing Intercept Wild - - 0.82 Foster -14.86 (2284.1) 0.99 Puppet -0.10 (0.86) 0.91 Nest site Intercept North Island - - 0.46 South Island 0.75 (1.10) 0.50 Nest site Intercept Insurance - - 0.44 Breeding 0.96 (0.95) 0.31 Wild -0.18 (1.03) 0.86 Location Intercept North Island - - 0.63 South Island 0.54 (1.10) 0.65 Location Intercept Insurance - - 0.40 Breeding -1.30 (1.14) 0.26 Wild -0.85 (0.90) 0.34 Location Intercept Offshore - - 0.33 Mainland -0.80 (0.79) 0.32 Translocation Intercept No - - 0.16 * Yes 1.11 (0.79) 0.16 In degree Intercept High - - 0.19 * Low 1.02 (0.79) 0.20 Out degree Intercept High - - 0.23 * Low 1.18 (1.10) 0.28 Betweenness Intercept High - - 0.51 Low 0.76 (1.14) 0.51 Medium 1.42 (1.26) 0.26 Location Intercept Close to farms - - 0.65 Remote from farms 0.38 (0.86) 0.66 Nest site Intercept Close to farms - - 0.30 Remote from farms 0.85 (0.80) 0.29 Sampling period Intercept Period 1 (Mar-Apr 2012) - - 0.56 Period 2 (Aug-Nov 2012) -0.01 (2306.1) 0.99 Period 3 (Feb-Apr 2013) -0.39 (0.86) 0.70 *variable was included in a multivariate model
Appendix 9.3-7 Table of unidentified Campylobacter spp. univariate analyses
177
Variable Level Coefficient (SE) p-value ChiSq p-value Sex Intercept Female - - 0.92 Male 0.04 (0.38) 0.92 Age Intercept Adult - - 0.04 * Juvenile 0.92 (0.45) 0.04 Rearing Intercept Wild - - 0.07 * Foster 16.72 (1385.38) 0.99 Puppet 0.44 (0.41) 0.28 Nest site Intercept North Island - - 0.75 South Island -0.13 (0.42) 0.75 Nest site Intercept Insurance - - 0.23 * Breeding 0.83 (0.52) 0.11 Wild 0.11 (0.42) 0.26 Location Intercept North Island - - 0.75 South Island 0.15 (0.45) 0.75 Location Intercept Insurance - - 0.63 Breeding 0.44 (0.47) 0.35 Wild 0.29 (0.44) 0.51 Location Intercept Offshore - - 0.05 * Mainland 0.85 (0.43) 0.05 Translocation Intercept No - - 0.90 Yes 0.05 (0.39) 0.90 In degree Intercept High - - 0.36 Low -0.36 (0.39) 0.36 Out degree Intercept High - - 0.52 Low -0.25 (0.39) 0.52 Betweenness Intercept High - - 0.50 Low -0.18 (0.42) 0.68 Medium -0.69 (0.59) 0.25 Location Intercept Close to farms - - 0.11 * Remote from farms 0.61 (0.39) 0.12 Nest site Intercept Close to farms - - 0.01 * Remote from farms 1.08 (0.45) 0.02 Sampling period Intercept Period 1 (Mar-Apr 2012) - - - * Period 2 (Aug-Nov 2012) 0.18 (0.77) 0.86 Period 3 (Feb-Apr 2013) -2.55 (0.42) 0.01 *variable was included in a multivariate model
Appendix 9.3-8 Univariate analyses of variables in respect to multiple carriage of Campylobacter spp.
178
9.4. Chapter 4 supplementary information
See excel spreadsheet for tables of factors, cofactors and allelic profiles per C. sp. nova 1 genome.
Additional sheets contain distance matrices for: Campylobacter sp. nova 1 allele, uncorrected P
distance and GTR distance matrices, and the uncorrected P distance matrix for the combined C. sp.
nova 1 and publically available Campylobacter spp. rMLST analysis.
Appendix 9.4-1 Electronic supplementary material
179
Campylobacter species Strain GenBank assembly ID Campylobacter coli 15-537360 GCA_000494775.1 Campylobacter coli 76339 GCA_000470055.1 Campylobacter coli RM1875 GCA_000583755.1 Campylobacter coli RM4661 GCA_000583775.1 Campylobacter coli CVM N29710 GCA_000465235.1 Campylobacter coli 1091 GCA_000253675.2 Campylobacter coli 1098 GCA_000253695.2 Campylobacter coli 1417 GCA_000253735.2 Campylobacter coli 1957 GCA_000253875.2 Campylobacter coli 2680 GCA_000253515.2 Campylobacter coli 111-3 GCA_000253415.2 Campylobacter coli 151-9 GCA_000254095.2 Campylobacter coli 84-2 GCA_000253615.2 Campylobacter coli 202-04 GCA_000253915.2 Campylobacter coli H9 GCA_000254195.2 Campylobacter coli JV20 GCA_000146835.1 Campylobacter coli K3 GCA_000505605.1 Campylobacter coli RM2228 GCA_000167415.1 Campylobacter coli RM5611 GCA_000583795.1 Campylobacter coli Z163 GCA_000253455.2 Campylobacter concisus 13826 GCA_000017725.1 Campylobacter concisus UNSWCD GCA_000259315.1 Campylobacter concisus UNSW3 GCA_000466645.1 Campylobacter concisus UNSW1 GCA_000466665.1 Campylobacter concisus UNSWCS GCA_000466685.1 Campylobacter concisus ATCC51561 GCA_000466705.1 Campylobacter concisus UNSW2 GCA_000466725.1 Campylobacter concisus ATCC51562 GCA_000466745.1 Campylobacter cuniculorum DSM23162 GCA_000621005.1 Campylobacter curvus 525.92 GCA_000017465.1 Campylobacter curvus DSM6644 GCA_000376325.1 Campylobacter fetus subsp. fetus 82-40 GCA_000015085.1 Campylobacter fetus subsp. testudinum 03-427 GCA_000495505.1 Campylobacter fetus subsp. venerealis Azul-94 GCA_000174675.1 Campylobacter fetus subsp. venerealis NCTC 10354 GCA_000222425.1 Campylobacter fetus subsp. venerealis bv. Intermedius 99541 GCA_000414135.1 Campylobacter fetus subsp. venerealis bv. Intermedius cfvi03/293 GCA_000512745.1 Campylobacter gracilis RM3268 GCA_000175875.1 Campylobacter hominis BAA-381 GCA_000017585.1 Campylobacter hyointestinalis subsp. hyointestinalis DSM_19053 GCA_000705275.1 Campylobacter jejuni subsp. doylei 49349 GCA_000701745.1 Campylobacter jejuni subsp. doylei 269.97 GCA_000017485.1 Campylobacter jejuni subsp. jejuni RM1221 GCA_000011865.1 Campylobacter jejuni subsp. jejuni NCTC11168 GCA_000009085.1 Campylobacter jejuni subsp. jejuni 00-2425 GCA_000468915.1
180
Campylobacter species Strain GenBank assembly ID Campylobacter jejuni subsp. jejuni 4031 GCA_000493495.1 Campylobacter jejuni subsp. jejuni 84-25 GCA_000168195.1 Campylobacter jejuni subsp. jejuni CF93-6 GCA_000168115.1 Campylobacter jejuni subsp. jejuni HB93-13 GCA_000168175.1 Campylobacter jejuni subsp. jejuni D2600 GCA_000234545.1 Campylobacter jejuni subsp. jejuni DFVF1099 GCA_000184805.2 Campylobacter jejuni subsp. jejuni 1336 GCA_000163975.1 Campylobacter jejuni subsp. jejuni IA3902 GCA_000025425.1 Campylobacter jejuni subsp. jejuni ICDCCJ07001 GCA_000184085.1 Campylobacter jejuni subsp. jejuni M1 GCA_000148705.1 Campylobacter jejuni subsp. jejuni NW GCA_000234525.1 Campylobacter jejuni subsp. jejuni R414 GCA_000163995.1 Campylobacter jejuni subsp. jejuni S3 GCA_000184205.1 Campylobacter jejuni subsp. jejuni PT14 GCA_000302555.2 Campylobacter jejuni subsp. jejuni 81-176 GCA_000015525.1 Campylobacter jejuni subsp. jejuni 81116 GCA_000017905.1 Campylobacter jejuni subsp. jejuni 10186 GCA_000686225.1 Campylobacter jejuni subsp. jejuni 260.94 GCA_000168135.1 Campylobacter jejuni subsp. jejuni ATCC33560 GCA_000251165.2 Campylobacter jejuni subsp. jejuni CG8486 13826 Campylobacter lari RM2100 GCA_000019205.1 Campylobacter mucosalis DSM21682 GCA_000705255.1 Campylobacter rectus RM3267 GCA_000174175.1 Campylobacter showae RM3277 GCA_000175655.1 Campylobacter showae CSUNSWCD GCA_000313615.1 Campylobacter showae CC57C GCA_000344295.1 Campylobacter upsaliensis RM3195 GCA_000167395.1 Campylobacter upsaliensis JV21 GCA_000185345.1 Campylobacter upsaliensis DSM5365 GCA_000620965.1 Campylobacter ureolyticus DSM20703 GCA_000374605.1
Campylobacter ureolyticus ACS-301-V-Sch3b GCA_000413435.1
Campylobacter ureolyticus CIT007 GCA_000597825.1 Appendix 9.4-2 List of publicly available Campylobacter spp. genomes used for rMLST comparison to Campylobacter sp. nova 1
181
Appendix 9.4-3 Multidimensional scaling plot of 70 Campylobacter sp. nova 1 52 gene rMLST haplotypic profiles, coloured by location of the takahe host
182
Factor d.f SS Estimate of variation Pseudo-F P(perm) Permutations Sampling location 5 4.99E-03 8.28E-05 6.4102 0.0001 9921 Cofactors North vs South Island 1 3.00E-04 9.12E-06 1.3932 0.2101 6984 Wild vs Captive breeding and Insurance 1 1.83E-03 5.08E-05 9.4558 0.0004 9958 Captive breeding vs Insurance 1 4.52E-04 1.01E-05 2.0075 0.1265 9947 Island vs Mainland 1 4.14E-04 7.25E-06 1.9389 0.1301 9958 Residuals 64 9.96E-03 1.56E-04 Total 69 1.50E-02
Appendix 9.4-4 PERMANOVA results from 9,999 permutations (perm) of location factors using an uncorrected P measure matrix from 52 rMLST nucleotide comparison of 70 takahe (Porphyrio hochstetteri) Campylobacter sp. nova 1
Factor d.f SS Estimate of variation Pseudo-F P(perm) Permutations Sampling location 5 1.63E-02 2.41E-04 4.0546 0.0001 9866 Cofactors North vs South Island 1 1.66E-03 7.47E-05 1.7148 0.1333 7053 Wild vs Captive breeding and Insurance 1 5.78E-03 1.52E-04 6.3545 0.0002 9922 Captive breeding vs Insurance 1 1.83E-03 3.68E-05 1.8267 0.0864 9931 Island vs Mainland 1 1.97E-03 3.62E-05 2.0374 0.0636 9922 Residuals 64 5.14E-02 8.03E-04 Total 69 6.77E-02
Appendix 9.4-5 PERMANOVA results from 9,999 permutations (perm) of location factors using a GTR matrix from 52 rMLST nucleotide comparison of 70 takahe (Porphyrio hochstetteri) Campylobacter sp. nova 1
183
9.5. Chapter 5 supplementary information
Appendix 9.5-1 Sequence alignment of the in silico PCR target region for Campylobacter sp. nova 1 isolates identified by in vitro PCR. Nucleotide polymorphisms which differ from the consensus sequence have been highlighted. Nucleotide bases are colour coded as follows: A = Yellow, T = Green, C = Blue, G = Yellow.
184
O what a joyous joyous day Is that on which we come At the recess from school away, Each lad to his own home!
What though the coach is crammed full, The weather very warm; Think you a boy of us is dull, Or feels the slightest harm?
The dust and sun is life and fun; The hot and sultry weather A higher zest gives every breast, Thus jumbled all together.
Sometimes we laugh aloud aloud, Sometimes huzzah, huzzah. Who is so buoyant, free, and proud, As we home-travellers are?
But sad, but sad is every lad That day on which we come, That last last day on which away We all come from our home.
The coach too full is found to be: Why is it crammed thus? Now every one can plainly see There’s not half room for us.
Soon we exclaim, O shame, O shame, This hot and sultry weather, Who but our master is to blame, Who pack’d us thus together!
Now dust and sun does every one Most terribly annoy; Complaints begun, soon every one Elbows his neighbour boy.
Not now the joyous laugh goes round, We shout not now huzzah; A sadder group may not be found Than we returning are.
---The Journey to and from school Mary Lamb
185
186